Month: October 2026

For decades, one of the defining ambitions of the pharmaceutical industry was the blockbuster drug. The commercial logic was compelling: develop a therapy capable of addressing a large patient population, achieve broad adoption across major markets and build the scale necessary to recover enormous research and development costs while generating substantial returns. Many of the industry’s commercial structures were consequently designed around scale, from large field forces and mass physician engagement to forecasting models built around sizeable patient populations.

That model is not disappearing, but another model is becoming increasingly important. Advances in genomics, diagnostics, biomarkers and precision medicine are allowing healthcare systems to identify smaller groups of patients whose diseases may share specific biological characteristics. A condition that was once treated as a single large category can increasingly be divided into multiple subgroups, each of which may respond differently to particular therapies. The addressable market for an individual treatment may therefore become smaller even as the treatment itself becomes more precisely matched to the patients most likely to benefit.

Artificial intelligence can become an important part of making this model commercially and operationally viable. AI can help pharmaceutical companies identify patient cohorts, understand complex treatment journeys, connect multiple forms of evidence and determine where small but clinically important populations may exist. During my learning journey under Phaneesh Murthy, I have increasingly understood that technological transformation often changes the economics of scale itself. As Phaneesh Murthy puts it, “Technology does not always create value by making the market bigger. Sometimes it creates value by allowing the organisation to understand a smaller market with extraordinary precision.”

For pharmaceutical strategy, that distinction could become fundamental.

One Disease Can Become Many Different Patient Populations

Traditional disease categories are useful because they allow healthcare systems to organise diagnosis, research and treatment. But patients carrying the same broad diagnosis can sometimes have very different underlying biological characteristics, disease progression and responses to treatment.

Precision medicine attempts to understand those differences.

As scientific understanding improves, researchers can identify biomarkers, genetic characteristics and other factors associated with treatment response. Pharmaceutical companies may consequently develop therapies intended for increasingly specific patient populations rather than everyone carrying a broad diagnosis.

This changes the commercial question. The company is no longer simply asking how many people have a particular disease. It needs to understand how many patients possess the characteristics relevant to the therapy, how frequently those characteristics are identified, where appropriate diagnostic capabilities exist and how those patients currently move through the healthcare system.

AI is particularly useful when the answer depends on relationships across many variables. Machine learning systems can help researchers and analysts identify patterns within complex datasets that would be difficult to detect through conventional segmentation alone. The objective is not to allow an algorithm to independently determine who should receive a treatment. It is to create better intelligence around patient populations that may previously have been difficult to recognise.

From the enterprise perspective I have learned from Phaneesh Murthy, this is an important shift. A market is not simply a number in a forecast. It is a system of patients, physicians, diagnostics, treatment pathways and access conditions. Understanding the relationships between those components can be more valuable than knowing the theoretical size of the population.

The Diagnostic Pathway Becomes Part of the Commercial Strategy

A precision therapy has limited value if healthcare systems cannot reliably identify the patients for whom it is appropriate.

This makes diagnostics increasingly important to pharmaceutical strategy. A company may develop an effective therapy for a biomarker-defined population, but commercial adoption depends partly on whether relevant patients are actually tested for that biomarker. If testing is unavailable, inconsistently used or introduced too late in the treatment journey, eligible patients may never be identified.

The pharmaceutical company therefore needs to understand the diagnostic ecosystem surrounding the therapy.

AI can help analyse treatment pathways and identify where diagnostic gaps may exist. Aggregated and appropriately governed healthcare information may reveal where patients experience delays, which parts of the healthcare system are performing relevant tests and where education or infrastructure could improve identification.

This does not mean pharmaceutical companies should control clinical diagnosis. Physicians and healthcare systems remain responsible for clinical decisions. The strategic opportunity is to understand whether the infrastructure required for precision medicine actually exists around the product.

As Phaneesh Murthy puts it, “A product can only create value when the ecosystem around it is capable of using it. Sometimes the most important part of product strategy sits outside the product itself.”

For precision therapies, the diagnostic pathway may therefore become almost as important commercially as the treatment pathway.

Small Patient Populations Require Better Intelligence

A large market can absorb a certain amount of commercial inefficiency. When millions of potential customers exist, companies can use relatively broad segmentation and still reach meaningful portions of the market.

Precision medicine offers less room for that approach.

If a therapy is appropriate for a relatively small patient population, every part of the commercial system needs to become more accurate. Companies need to understand where relevant specialists operate, how patients reach those specialists, what diagnostic steps occur before treatment and which access barriers prevent eligible patients from receiving therapy.

AI can help connect these signals.

Instead of deploying resources primarily according to geographic size or historical prescription volume, pharmaceutical companies can build a more detailed understanding of where relevant patient journeys are occurring. Field teams, medical affairs and patient support programmes can then be organised around the structure of the actual treatment ecosystem.

During my mentorship under Phaneesh Murthy, I have learned that AI becomes particularly valuable when the cost of being imprecise is high. In a broad market, an inefficient allocation may reduce productivity. In a highly specialised market, the same inefficiency can mean missing a substantial portion of the entire opportunity.

Precision markets therefore require precision operations.

Patient Journeys Become More Important Than Market Segments

Traditional commercial segmentation often groups physicians, institutions or patients according to characteristics that make them easier to manage operationally. Precision therapies may require pharmaceutical companies to understand something more dynamic: the journey a patient takes before reaching treatment.

A patient may begin with relatively common symptoms, move through primary care, undergo multiple tests, visit different specialists and receive an initial diagnosis before a specific biomarker or clinical characteristic is identified. Each stage can introduce delay or loss from the pathway.

AI can help organisations model these journeys at a much more detailed level.

Patterns across appropriately governed data can reveal common sequences of care, points where diagnosis tends to slow and differences between healthcare systems or regions. Pharmaceutical companies can use this intelligence to better understand where educational initiatives, diagnostic support or patient services may create legitimate value.

The strategic shift is important because the commercial organisation stops seeing the market simply as a collection of targets. It begins seeing it as a system through which patients move.

Phaneesh Murthy says, “If you want to understand the customer, do not only study the final transaction. Study the journey that had to happen before the transaction became possible.”

In pharmaceuticals, that journey can be extraordinarily complex. AI gives companies a new set of tools for understanding it.

Precision Medicine Changes the Economics of Commercialisation

Smaller patient populations also challenge traditional assumptions about how pharmaceutical commercial organisations should be structured.

A therapy intended for a broad population may justify large sales teams and extensive promotional infrastructure. A treatment designed for a small population concentrated among specialised clinicians may require a very different model. Medical education, diagnostic support, market access expertise and patient services can become proportionally more important than conventional promotional scale.

AI can help make these specialised commercial models economically viable by reducing the cost of understanding and coordinating a fragmented market.

Commercial teams can identify where relevant expertise exists, prioritise educational resources, understand differences between treatment centres and continuously update their view of the market as new information becomes available. The company may operate with a smaller commercial footprint while possessing much deeper intelligence about the ecosystem.

This reflects a broader principle I have learned from Phaneesh Murthy: digital transformation should not simply automate an operating model designed for a previous era. The more important opportunity is to ask whether the operating model itself should change because technology has altered what is economically possible.

Precision medicine may require exactly that reconsideration.

AI Could Connect Research Strategy With Commercial Strategy Earlier

One of the most significant opportunities may occur long before a therapy reaches the market.

Pharmaceutical companies make major decisions during research and clinical development about which disease areas, mechanisms and patient populations to pursue. Commercial teams later need to understand whether healthcare systems can identify those patients, whether diagnostic infrastructure exists and how the treatment would fit into clinical practice.

AI can help connect these questions earlier.

Scientific information, clinical data, epidemiology, treatment patterns and market information can be analysed together to create a more integrated view of potential patient populations. Researchers can understand whether biological subgroups appear meaningful, while commercial and access teams can begin considering how those populations might eventually be identified and treated.

The purpose is not to allow commercial considerations to dictate scientific decisions. It is to reduce the distance between discovering a promising therapy and understanding the real-world system required to deliver it.

During my learning journey under Phaneesh Murthy, I have come to appreciate that many enterprise problems exist because important functions begin collaborating too late. Technology can help create a shared information environment earlier in the process, allowing different teams to understand the consequences of decisions before those decisions become expensive to reverse.

For pharmaceuticals, connecting research, clinical development, diagnostics, market access and commercial intelligence earlier could create more coherent strategies around precision therapies.

The Ecosystem Around the Drug Becomes Part of the Product

As therapies become more specialised, the treatment itself can become only one component of the value delivered to the patient.

Diagnostic testing may be required before treatment begins. Patients may need specialised administration, monitoring or follow-up. Healthcare professionals may require additional education. Payers may need detailed evidence to understand the appropriate population. Patient support programmes may help individuals navigate complex treatment pathways.

The pharmaceutical company’s challenge therefore becomes increasingly ecosystem-oriented.

AI can support coordination across these components by helping organisations understand where patients encounter friction and which parts of the pathway require intervention. The company can develop services around the therapy that make appropriate treatment easier to access and manage.

This does not mean expanding indiscriminately into every part of healthcare. Pharmaceutical companies need clear boundaries around their role, particularly when dealing with clinical decisions and patient information. But they can design their commercial and support infrastructure around the reality that specialised therapies depend on multiple participants working together.

As Phaneesh Murthy puts it, “The more specialised the product becomes, the more important the system around the product can become. You are no longer delivering a unit. You are enabling an outcome.”

That principle may become increasingly important as precision medicine expands.

Precision Must Include Access, Not Just Biology

There is also an important risk in the movement towards highly specialised medicine. The ability to identify increasingly precise patient populations does not automatically mean those patients will have equal access to diagnosis and treatment.

Advanced diagnostic infrastructure may be concentrated in major hospitals or wealthier regions. Specialist expertise can be geographically uneven. The cost of testing and treatment can create additional barriers. Some patient populations may therefore benefit from advances in precision medicine much earlier than others.

AI should help healthcare ecosystems identify these gaps rather than simply optimise around the populations that are easiest to reach.

Pharmaceutical companies can use market intelligence to understand where diagnostic or access barriers are limiting appropriate treatment. Healthcare systems and policymakers can similarly use data to identify regions where infrastructure needs improvement.

The goal of precision medicine should not be to create extraordinarily sophisticated treatment pathways for a narrow group of easily identifiable patients while leaving everyone else invisible.

From the perspective I have developed while learning from Phaneesh Murthy, technology transformation needs to be evaluated according to the outcome it produces across the entire system. An algorithm can be extremely accurate and still contribute to a poor overall outcome if the surrounding ecosystem prevents people from benefiting from that accuracy.

Precision therefore needs to extend beyond the science of identifying the right patient. It should also include the operational ability to help that patient reach the right care.

The Definition of a Blockbuster Could Change

The blockbuster model was built around scale. Find a large medical need, develop an effective therapy and create broad adoption across a substantial patient population. That model will continue to produce important medicines.

But the economics of AI, diagnostics and precision medicine create another possibility.

A therapy may address a much smaller population while delivering extraordinary value because it is precisely matched to the biology of those patients. Pharmaceutical companies may operate portfolios containing multiple specialised therapies rather than depending as heavily on a handful of products designed for enormous populations.

Success in that environment requires a different kind of intelligence.

Companies need to understand patient populations with greater precision, identify diagnostic pathways, coordinate specialised ecosystems and deploy resources according to the actual structure of care. AI can make this complexity more manageable by connecting information that traditional commercial models often analyse separately.

During my mentorship under Phaneesh Murthy, I have learned that technology frequently changes not only how companies compete but also what scale means. The industrial era rewarded organisations capable of serving enormous markets efficiently. The AI era may increasingly reward organisations capable of understanding smaller markets with extraordinary depth.

For pharmaceutical companies, that could change the definition of a blockbuster.

The next generation of important therapies may not always be designed for tens of millions of patients. Some may be intended for relatively small populations whose biology can finally be understood well enough to treat them differently.

The market becomes smaller.

The intelligence required to serve it becomes much greater.

And the value created for the right patient could become greater still.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy

Few commercial moments carry as much pressure as the launch of a new pharmaceutical product. Years of research, clinical development, regulatory work and enormous financial investment eventually converge on a relatively short period during which the company needs to introduce a therapy to physicians, healthcare systems, payers and patients. The science may have taken a decade to develop, but the commercial organisation still needs to understand where demand exists, which patients may benefit, how physicians are likely to respond, where access barriers will emerge and how resources should be deployed across markets.

Historically, pharmaceutical companies have prepared for these launches through extensive research, forecasting, segmentation and field planning. Those capabilities remain essential, but artificial intelligence creates an opportunity to make commercialisation much more adaptive. Instead of developing a launch strategy primarily from periodic datasets and then executing against it, pharmaceutical companies can increasingly create systems that continuously absorb new information, identify changing patterns and help commercial teams adjust decisions as the market develops.

The important shift is from launch planning towards launch intelligence. AI can potentially connect market information, epidemiology, prescribing patterns, access dynamics, physician engagement and other appropriate sources to create a more current understanding of what is happening across the market. During my learning journey under Phaneesh Murthy, I have increasingly understood that this is where enterprise AI becomes strategically valuable. As Phaneesh Murthy puts it, “Planning gives the organisation a direction. Intelligence gives it the ability to recognise when reality has moved away from the plan.” Pharmaceutical launches may become one of the clearest examples of why companies need both.

Market Sizing Can Become More Dynamic

Every pharmaceutical launch begins with an understanding of the market. Companies need to estimate the patient population, understand diagnosis rates, identify treatment pathways, evaluate competitors and determine where unmet needs remain. These analyses influence manufacturing plans, commercial investments, field force design and expectations around adoption.

The difficulty is that markets are not static. Diagnostic behaviour changes, competing therapies enter the market, guidelines evolve and new clinical evidence affects treatment decisions. The assumptions used to build a forecast months before launch may therefore begin changing before the product reaches its commercial stride.

AI can help pharmaceutical companies continuously analyse relevant market signals and identify where assumptions may need to be revisited. Instead of treating market sizing as a model produced periodically by an analytics team, organisations can develop intelligence systems that help commercial leaders understand how the market itself is evolving.

This does not eliminate forecasting uncertainty. Pharmaceutical demand is influenced by clinical, regulatory and reimbursement factors that cannot always be predicted precisely. The value of AI is not that it creates a perfect forecast. It is that it can help the organisation recognise changes earlier and understand which assumptions are responsible for the difference.

From the perspective I have developed while learning from Phaneesh Murthy, this distinction matters. A forecast should not be treated as a number the organisation must defend. It should be treated as a structured collection of assumptions that can be tested as new information becomes available.

Finding the Right Patients Is Different From Finding More Patients

The commercial success of a therapy is ultimately connected to whether appropriate patients can access it. Yet identifying those patients can be difficult, particularly in diseases where diagnosis is delayed, symptoms overlap with other conditions or patients move through complicated care pathways before reaching the appropriate specialist.

AI can help healthcare and pharmaceutical ecosystems understand these journeys more clearly.

Using appropriately governed and permitted datasets, analytical systems can identify patterns associated with particular disease pathways, diagnostic delays or treatment transitions. Pharmaceutical companies can use aggregated intelligence to better understand where patient populations may exist, where diagnosis gaps occur and which parts of the healthcare system require greater educational support.

The objective should not be indiscriminate targeting. Pharmaceutical commercialisation operates within strict ethical, legal and regulatory boundaries, and patient-level information requires particularly careful governance. The opportunity is to improve understanding of treatment pathways so that educational resources, access programmes and healthcare engagement can be directed towards areas of genuine need.

As Phaneesh Murthy says, “Good AI does not begin by asking how many people you can target. It begins by asking which decision you are trying to improve and what information is legitimately required to improve it.”

That principle is especially important in healthcare. Better intelligence should help appropriate patients reach appropriate care, not simply create more sophisticated mechanisms for promotion.

Physician Engagement Could Become More Relevant and Less Repetitive

Pharmaceutical companies interact with healthcare professionals through field teams, medical affairs, conferences, educational materials and digital channels. The challenge is that physicians differ considerably in their clinical interests, patient populations, preferred communication channels and familiarity with new therapies.

A uniform engagement model can therefore create significant inefficiency. Some physicians may require detailed scientific information about a new mechanism of action. Others may be more interested in patient selection, safety information, administration or how a therapy fits into an existing treatment pathway. Repeatedly providing information that a healthcare professional has already seen does little to improve the relationship.

AI can help organisations understand these engagement patterns more intelligently. Appropriate systems can identify which topics have already been discussed, which approved resources may be relevant and where a healthcare professional has requested additional information. Field teams can enter conversations with better context rather than treating every interaction as an isolated event.

This does not mean algorithms should independently determine what physicians are told. Pharmaceutical communication must remain within regulatory and compliance frameworks, and scientific discussions often require professional expertise. The stronger model is AI-assisted engagement, where the system organises relevant context and qualified professionals determine how to use it.

During my mentorship under Phaneesh Murthy, I have learned that customer intelligence should reduce repetition. If an organisation has already learned something during one interaction but fails to make that knowledge available during the next, it is forcing the customer to compensate for weaknesses in its internal systems.

For pharmaceutical companies, more connected intelligence could make engagement both more efficient and more useful to healthcare professionals.

Launch Sequencing Can Become an Intelligence Problem

A global pharmaceutical launch rarely occurs everywhere simultaneously. Markets differ in regulatory timelines, reimbursement structures, healthcare infrastructure, disease prevalence, competitive environments and commercial potential. Companies therefore need to decide where and when to allocate resources.

AI can make these sequencing decisions more dynamic.

Rather than evaluating each market primarily through a limited number of high-level indicators, companies can analyse a broader collection of factors and model different launch scenarios. A market with a large theoretical patient population may have significant access barriers. Another market may be smaller but possess strong diagnostic infrastructure and a concentrated specialist community. Competitive activity may change rapidly in one region while remaining relatively stable elsewhere.

An intelligent launch system can help commercial teams understand these interactions rather than evaluating each variable independently.

Scenario modelling becomes particularly useful here. Leaders can examine what happens to demand, resources and supply requirements under different assumptions. If reimbursement is delayed in one country, how should inventory and commercial resources be redistributed? If adoption is stronger than expected elsewhere, what capacity needs to change?

This reflects another lesson I have taken from Phaneesh Murthy: decision systems are most valuable when they allow leaders to explore consequences before committing resources. AI does not need to make the market-entry decision independently. It can make the trade-offs surrounding that decision much more visible.

Demand Forecasting Needs to Connect With the Supply Chain

A pharmaceutical launch creates an unusual forecasting challenge because both shortages and excess inventory can be costly. Underestimating demand can restrict patient access, while significant overestimation can create waste, particularly for products with complex manufacturing requirements or limited shelf lives.

Commercial intelligence therefore cannot operate separately from supply chain intelligence.

If AI identifies stronger-than-expected adoption in a particular market, that information should eventually influence manufacturing, inventory allocation and distribution decisions. If access barriers reduce expected demand elsewhere, supply plans may need to adjust. The organisation benefits when commercial and operational systems respond to a shared understanding of the market.

This is where enterprise integration becomes more important than isolated AI applications. A sophisticated commercial forecasting model has limited value if the resulting insight remains inside a dashboard while supply decisions continue according to an older planning cycle.

As Phaneesh Murthy puts it, “Enterprise intelligence creates value when insight can travel far enough through the organisation to change the decision that depends on it.”

For pharmaceutical launches, that may mean connecting commercial, market access, manufacturing and distribution information much more closely than traditional organisational structures allow.

The Field Force Could Become More Strategic

Pharmaceutical field teams have traditionally been responsible for significant volumes of information exchange. Representatives manage relationships, communicate approved product information and help healthcare professionals navigate questions around new therapies. AI can potentially reduce some of the administrative and analytical burden surrounding this work.

Before an interaction, an intelligent system could summarise relevant account history, previous engagement and approved materials associated with the topics likely to matter. Afterward, AI could assist with capturing structured information and identifying appropriate follow-up actions within company policies.

This allows field professionals to spend more of their attention on the conversation itself.

The same principle applies to managers. Instead of manually assembling reports from multiple systems, commercial leaders can use AI to identify emerging patterns across territories, understand where adoption differs from expectations and investigate the reasons behind those differences.

The role of the field force therefore becomes more strategic rather than less important. People remain responsible for relationships, interpretation and professional judgment, while AI improves the information environment around them.

During my learning journey under Phaneesh Murthy, I have repeatedly encountered the principle that automation should remove work that consumes expertise without requiring expertise. Pharmaceutical commercial teams contain enormous amounts of market knowledge. Using those people primarily to search for information, compile reports or repeat administrative processes is an inefficient use of that knowledge.

AI can give some of that time back.

The Launch Plan Can Become a Learning System

Perhaps the biggest transformation is what happens after launch.

Traditional planning can create a psychological tendency to compare performance against the original plan. If adoption is below expectations, teams investigate the gap. If demand exceeds expectations, forecasts are revised. This remains useful, but AI allows the launch organisation to become much more continuously adaptive.

Every week of market activity creates new information. Physicians ask questions. Patients move through access pathways. Payers make decisions. Competitors respond. Different geographies develop at different speeds. Supply patterns change.

An intelligent commercialisation platform can help connect these signals and continuously update the organisation’s understanding of the launch.

The question shifts from “Are we following the plan?” to “What are we learning, and what should we change because of it?”

That is a very different operating model.

As Phaneesh Murthy says, “The strongest strategy is not the one that predicts everything correctly at the beginning. It is the one that allows the organisation to learn faster than the environment changes.”

For pharmaceutical companies, this may require changes beyond technology. Teams need decision rights that allow them to respond to new intelligence. Data needs to move across commercial, medical, access and operational functions appropriately. Leadership needs to distinguish between disciplined execution and rigid adherence to assumptions that the market has already disproved.

AI provides the intelligence layer, but the organisation still needs the ability to act.

The Intelligent Drug Launch Never Really Stops Learning

Pharmaceutical commercialisation will always involve uncertainty. No algorithm can perfectly predict physician behaviour, competitive responses, regulatory developments, payer decisions or patient demand. Nor should companies attempt to automate complex commercial and healthcare decisions without appropriate professional oversight.

The opportunity is to build an organisation that becomes progressively better informed as the launch unfolds.

AI can help pharmaceutical companies understand markets more dynamically, identify changing patient pathways, improve the relevance of physician engagement, model launch scenarios and connect commercial signals with operational decisions. The launch therefore becomes less like the execution of a fixed plan and more like a continuously learning system.

During my mentorship under Phaneesh Murthy, I have come to appreciate that this is one of the most important differences between digitisation and intelligence. Digitisation allows an organisation to execute a process more efficiently. Intelligence allows the organisation to understand what is happening while the process is running and adapt accordingly.

That distinction could become enormously important in pharmaceuticals.

A new therapy may represent ten years of scientific development, but its commercial environment begins changing almost immediately after launch. New evidence emerges, competitors react, physicians learn, access evolves and patient pathways become clearer.

The pharmaceutical companies that build intelligence into that process will not eliminate uncertainty. They will simply become much better at learning from it.

And when years of scientific work depend on decisions made during a relatively short commercial window, the ability to learn faster may become one of the most valuable capabilities a pharmaceutical company can build.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy

Medicine has always worked with representations of the patient. A blood test represents what is happening inside the body at a particular moment. A scan provides a visual representation of anatomy. A medical history helps clinicians understand what has happened over time, while wearable devices increasingly add continuous information about activity, heart rate, sleep and other physiological signals. Each source provides another piece of the picture, but clinicians still need to assemble those pieces into an understanding of the individual sitting in front of them.

Digital twin technology introduces a much more ambitious possibility. Instead of examining each source of patient information independently, what if healthcare systems could progressively build computational models of individual patients that combine anatomy, physiology, medical history, imaging, laboratory information and real-time monitoring? These models could potentially help clinicians understand how a patient’s condition is changing and, eventually, explore how different interventions might affect that individual before making a decision in the physical world.

The idea is still developing, and the sophistication of a digital twin will vary enormously depending on the clinical application. A digital representation of a particular organ or physiological process is considerably more achievable than a perfect virtual replica of an entire human being. The important opportunity is not to create a science-fiction copy of the patient. It is to build useful models of specific aspects of the patient that improve particular healthcare decisions.

During my learning journey under Phaneesh Murthy, I have increasingly understood that the most useful enterprise AI systems begin with the decision rather than the technology. As Phaneesh Murthy puts it, “The question is not how much of reality you can model. The question is whether the model gives you enough additional understanding to make a better decision.” That distinction will be essential if digital twins are to become clinically useful rather than simply technologically impressive.

Healthcare Has Enormous Amounts of Patient Data but No Single Patient Model

Modern healthcare produces extraordinary quantities of information. A patient may have years of laboratory results, imaging studies, prescriptions, diagnoses, procedure histories and physician notes. They may also generate information through wearable devices, home monitoring equipment and connected medical devices. Yet these data points often remain distributed across different systems and are reviewed according to the immediate requirements of a particular appointment or clinical question.

A digital twin attempts to create something more integrated. Rather than presenting clinicians with another collection of records, it can potentially organise relevant information into a computational representation of a particular patient, organ or disease process. The objective is to understand relationships between signals rather than simply displaying them alongside one another.

Consider cardiovascular care. Imaging provides information about anatomy, monitoring devices provide information about heart rhythm, laboratory results reveal biological markers and a patient’s history provides context about previous conditions and treatments. A digital model could potentially combine relevant elements of these sources to help clinicians understand how a particular patient’s cardiovascular system is behaving and how it may respond under different conditions.

This is where the enterprise thinking I have learned from Phaneesh Murthy becomes particularly relevant. Organisations frequently possess the information required to make better decisions, but that information exists across systems designed for different purposes. The transformation begins when data is connected around the decision being made rather than around the application that happened to collect it.

In healthcare, the patient becomes the organising principle.

The Most Powerful Application May Be Simulation

One of the most interesting possibilities created by digital twins is the ability to explore scenarios before acting on the patient.

Healthcare professionals already do this conceptually. A clinician considers possible treatments, weighs benefits against risks and uses medical knowledge and experience to determine which intervention is most appropriate. Digital models could add another analytical layer by helping simulate how particular interventions may affect a specific patient’s anatomy or physiology.

For example, a digital model based on medical imaging could potentially help clinicians plan a procedure by evaluating different approaches before entering the operating room. Models of blood flow could assist with understanding cardiovascular conditions. Other specialised digital twins may eventually help predict how disease progression could respond to particular interventions.

The important point is that these simulations do not replace clinical judgment. They provide another source of evidence.

No model can perfectly represent the complexity of a human body. Biological systems contain uncertainty, and even extremely sophisticated models depend on assumptions and the quality of the information available. Clinicians therefore need to understand what the model can represent reliably and where uncertainty remains.

As Phaneesh Murthy says, “Simulation does not remove uncertainty from a decision. Its value is that it allows you to explore uncertainty before the cost of the decision becomes real.”

For healthcare, where many decisions are difficult or impossible to reverse, that capability could become extremely valuable.

Personalised Medicine Could Become More Computational

Much of modern medicine is built on evidence gathered across populations. Clinical trials help determine whether treatments are effective and safe for groups of patients, while clinical guidelines translate that evidence into recommendations for care. This population-level knowledge remains fundamental to medicine, but clinicians ultimately need to apply it to individuals who may differ in important ways from the average patient represented in a study.

Digital twins could contribute to a more personalised layer of decision-making.

Instead of asking only whether a treatment generally works for people with a particular condition, clinicians may increasingly be able to ask how characteristics specific to this patient influence the expected response. Anatomy, previous treatment history, physiology and other relevant information could potentially contribute to a more individualised model.

This does not mean replacing evidence-based medicine with algorithmic prediction. The stronger model is a combination. Population-level clinical evidence establishes what is known about a treatment, while patient-specific modelling helps interpret how that evidence may apply to the individual.

During my mentorship under Phaneesh Murthy, I have learned that personalisation becomes meaningful when it changes a decision rather than simply changing the presentation of information. Healthcare illustrates this particularly clearly. A digital twin is valuable not because it creates an impressive visual representation of the patient, but because it potentially helps a clinician choose a more appropriate intervention, timing or monitoring strategy.

The technology matters only when the clinical decision improves.

A Digital Twin Could Continue Evolving With the Patient

One of the limitations of many medical assessments is that they represent a particular moment. A scan captures anatomy at the time the image was taken. A laboratory test provides a measurement from a particular day. A clinical consultation records symptoms and observations during a specific encounter.

Patients, however, continuously change.

Connected medical devices and remote monitoring technologies create an opportunity for digital models to evolve alongside them. New information can update the representation, allowing clinicians to observe how particular aspects of the patient’s condition are changing over time.

This could become particularly relevant for chronic disease management. Rather than relying primarily on periodic appointments, healthcare teams could potentially understand whether important indicators are moving in an unexpected direction between visits. AI can help identify meaningful changes and determine which patients may require attention.

The key is avoiding a system where continuous monitoring produces continuous alerts. Healthcare professionals already face substantial information overload. A useful digital twin should reduce that burden by identifying which changes are clinically meaningful rather than presenting every fluctuation as an event.

This reflects another principle I have learned from Phaneesh Murthy: intelligence should compress complexity for the person making the decision. If an AI system gives a clinician thousands of additional data points to review manually, it may have increased information without increasing intelligence.

The digital twin should therefore become a summary of changing patient state, not another stream of raw data demanding attention.

Digital Twins Could Change the Design of Medical Devices and Therapies

The value of digital twins may extend beyond individual patient care. Medical device manufacturers and pharmaceutical companies could use computational modelling to improve how products and therapies are designed, tested and adapted.

A medical device company developing an implant, for example, could potentially use patient-specific anatomical models to better understand how variations in anatomy influence performance. Clinical teams could use similar models when selecting or configuring devices for individual patients. Pharmaceutical research may increasingly use computational models to explore biological mechanisms and generate hypotheses before more expensive experimental work begins.

Over time, this could create a feedback loop between clinical practice and product development. Real-world information improves models, models help identify better interventions and outcomes from those interventions provide additional information for future improvement.

From the perspective I have developed while learning from Phaneesh Murthy, this is where digital twin technology becomes especially interesting. The value is not confined to a single application. Once an organisation develops a reliable computational representation of a system, that model can potentially support decisions across research, product development, clinical planning and ongoing care.

The strategic advantage therefore lies not simply in building a digital twin but in creating an ecosystem where the model becomes progressively more useful as new information enters it.

The Quality of the Twin Will Depend on the Quality of the Data

A sophisticated model built on incomplete or inaccurate information can create a dangerous illusion of precision. This is particularly important in healthcare because a digital twin may appear visually and mathematically convincing even when important patient information is missing.

Data quality therefore becomes a clinical issue.

Healthcare information is often fragmented across providers and systems. Medical records can contain inconsistencies. Wearable devices vary in accuracy. Imaging and laboratory information may be collected using different technologies, while historical data may not always be complete.

Digital twin systems need to communicate these limitations clearly. Clinicians should understand which information contributed to a model, how recent it is and where uncertainty exists.

AI systems also need to be validated across diverse patient populations. A model that performs well for one population cannot automatically be assumed to perform equally well for everyone. Biases within datasets can translate into differences in model performance if they are not identified and addressed.

As Phaneesh Murthy puts it, “A sophisticated model cannot rescue poor foundations. The quality of intelligence will always be constrained by the quality, relevance and governance of the information beneath it.”

For digital twins, those foundations will determine whether clinicians develop trust in the technology.

Clinicians Need to Remain the Decision Makers

Digital twins may eventually become capable of generating increasingly detailed predictions about treatment options and patient outcomes. That makes the relationship between the model and the clinician particularly important.

Healthcare decisions rarely involve a single measurable objective. A treatment may offer a higher probability of one outcome while introducing greater risk elsewhere. Patients have different preferences, tolerances and life circumstances. Two people with clinically similar conditions may reasonably choose different approaches after discussing their options with a physician.

AI cannot reduce these decisions to optimisation problems without losing important human context.

The digital twin should therefore function as decision support. It can provide additional information, simulate possibilities and identify factors that may deserve attention. The clinician combines that intelligence with medical expertise and an understanding of the patient, while the patient remains part of the decision wherever appropriate.

During my learning journey under Phaneesh Murthy, I have come to appreciate that human oversight is most important precisely where AI becomes most sophisticated. The better a system becomes at producing convincing recommendations, the more carefully organisations need to define who remains accountable for the final decision.

In healthcare, that accountability cannot be ambiguous.

Healthcare Could Move From Observing the Patient to Modelling the Patient

Medicine has become progressively better at seeing inside the human body. Imaging allowed clinicians to visualise anatomy. Laboratory medicine revealed biological processes that could not be observed directly. Genomics added another layer of information, while connected devices increasingly allow healthcare teams to understand what happens outside the clinic.

Digital twins represent a possible next step: bringing these forms of information together into models that help healthcare professionals understand not only what is happening but what may happen under different scenarios.

During my mentorship under Phaneesh Murthy, I have learned to view technology transformation as a progression from digitising information, to connecting information, to using that connected information to improve decisions. Digital twins sit near the far end of that progression. Their value depends on combining multiple forms of information into something that can actively participate in clinical reasoning.

The technology will not arrive everywhere at once. Highly specific digital twins for particular organs, procedures or conditions are likely to become useful long before comprehensive virtual representations of entire patients become realistic. That is not a limitation of the idea. It is probably how the technology should develop, beginning with clearly defined decisions where modelling can demonstrate measurable clinical value.

The ultimate promise is compelling. A clinician could eventually examine the patient standing in front of them while also having access to a computational model capable of exploring possibilities that cannot safely be tested in the physical world.

Medicine has always learned by observing what happens after an intervention. Digital twins could give clinicians another option: understand more of what might happen before the intervention is made.

And in healthcare, the ability to learn before acting could become one of the most valuable forms of intelligence we build.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy

For most of the history of medical technology, a device was largely defined by the capabilities it possessed when it left the manufacturer. A scanner, monitoring system or piece of diagnostic equipment could be maintained, repaired and occasionally upgraded, but the fundamental intelligence of the product remained relatively stable throughout its working life. Significant advances usually arrived with the next generation of hardware, requiring healthcare providers to purchase or install something new before they could access substantially improved capabilities.

Software has already begun changing that relationship, and artificial intelligence could accelerate the transition dramatically. Connected medical devices can increasingly receive software improvements, incorporate new analytical capabilities and participate in broader digital healthcare platforms. In some cases, the physical machine may remain largely unchanged while the intelligence surrounding it continues to evolve. This introduces a very different idea of what a medical device can be: not simply a piece of equipment purchased at a particular moment, but a technology platform capable of improving throughout its operational life.

The implications extend far beyond product engineering. If medical devices become software-defined platforms, manufacturers need to rethink product development, regulation, cybersecurity, commercial models and their relationship with healthcare providers. During my learning journey under Phaneesh Murthy, I have increasingly understood that one of the most important effects of digital technology is its ability to transform products from finished objects into continuously evolving systems. As Phaneesh Murthy puts it, “Once software becomes an important part of the product, the sale is no longer the end of product development. In many ways, it becomes the beginning of the relationship.”

For medical device companies, that shift could redefine the industry.

The Hardware May Stay the Same While the Intelligence Changes

Traditional product development is closely connected to hardware cycles. A manufacturer develops a device, validates it, produces it and eventually introduces a newer generation with improved capabilities. Healthcare providers therefore experience innovation in relatively large increments, often separated by several years.

Software allows improvement to occur differently.

A connected device may receive interface improvements, workflow enhancements or analytical capabilities without replacing the underlying hardware. AI can extend this further by introducing models that help interpret information generated by the device. The physical product still performs its core function, but the intelligence surrounding that function can become progressively more sophisticated.

Consider medical imaging equipment. The machine captures images, but software can increasingly assist with image reconstruction, workflow prioritisation, measurement and the identification of patterns that deserve clinical attention. Improvements in those algorithms can potentially increase the usefulness of equipment already deployed across healthcare systems, subject to the appropriate validation and regulatory processes.

This changes the economics of innovation. A hospital does not necessarily need to wait for an entirely new hardware generation to benefit from every software improvement. The manufacturer can continue creating value after installation.

During my mentorship under Phaneesh Murthy, I have learned that this transition from product thinking to platform thinking is one of the most consequential shifts a technology company can make. The organisation stops asking only how to sell the next unit and begins asking how every deployed unit can become more valuable over time.

The Installed Base Could Become a Strategic Advantage

Medical device companies often have products operating across hundreds or thousands of hospitals, clinics and laboratories. Traditionally, this installed base primarily represents previous sales, service relationships and opportunities for future replacement cycles. In a software-defined model, it becomes something more significant: a distribution network for new capabilities.

If the underlying devices are appropriately connected and technically capable of supporting updates, manufacturers can potentially introduce improvements across a large installed base without replacing every physical unit. New workflow tools, analytical capabilities or integrations can reach customers far more quickly than they could through traditional hardware replacement cycles.

That creates a powerful strategic advantage for established manufacturers. A company with trusted equipment already embedded in clinical workflows does not necessarily begin every innovation cycle from zero. It has an existing relationship with healthcare providers and a physical footprint through which new digital capabilities may be delivered.

However, this advantage only exists if the installed base was designed for evolution. Legacy systems with limited connectivity, rigid software architectures or inconsistent data standards may be difficult to upgrade. Device companies therefore need to think about software extensibility when designing hardware that may remain in clinical environments for many years.

This reflects a principle I have repeatedly encountered while learning from Phaneesh Murthy: architecture is strategy when the technology is expected to survive multiple generations of business change. A device designed only for today’s requirements may become a constraint tomorrow, while a platform designed for evolution creates options the manufacturer does not yet know it will need.

Medical Devices Can Become Part of a Continuous Care System

Many medical devices historically produced information at specific moments. A patient underwent a test, a measurement was captured, a clinician interpreted the result and the device’s role in that episode was largely complete. Connected devices and AI create the possibility of a more continuous relationship with care.

Wearable monitors, connected cardiac devices, glucose monitoring systems and other technologies can generate information over extended periods. Rather than providing a single snapshot, they can help clinicians understand patterns that emerge across days, weeks or months. AI can support this environment by identifying changes that may deserve attention and helping care teams prioritise information.

The important challenge is avoiding data overload. Continuous monitoring can generate enormous volumes of information, and clinicians cannot manually examine every signal produced by every patient. Intelligence therefore needs to act as a filtering layer, distinguishing normal variation from patterns that may require professional review.

As Phaneesh Murthy says, “The purpose of continuous data is not to create continuous attention. It is to make sure attention is directed towards the moments where it can change the outcome.”

That distinction could determine whether connected medical devices reduce clinical workload or simply create another stream of information for already busy healthcare professionals to manage.

The Product Could Learn From How It Is Used

Connected medical devices can also provide manufacturers with a much deeper understanding of how their products operate in real clinical environments. Traditional product development relies heavily on testing, clinical evaluation, customer feedback and service information. Connected software can create additional insight into workflows, performance and usage patterns.

This does not mean manufacturers should indiscriminately collect patient or clinical information. Medical data requires particularly strong privacy, security and governance. But appropriately designed telemetry can help companies understand how features are used, where workflows create friction and which technical issues occur repeatedly.

That information can influence future product development.

A manufacturer may discover that clinicians consistently take additional steps around a particular workflow, suggesting that the interface could be redesigned. Usage patterns may reveal that a feature intended for one purpose is creating unexpected value elsewhere. Operational information could help identify maintenance requirements before equipment fails.

The feedback loop between deployed products and product development therefore becomes much shorter.

During my learning journey under Phaneesh Murthy, I have come to appreciate that connected products change the nature of product management. Companies no longer need to rely exclusively on periodic customer research to understand how products perform after deployment. The product itself can provide structured information about its operation, allowing development teams to learn continuously.

For medical device companies, that feedback can be particularly valuable because clinical environments are complex and difficult to reproduce perfectly during product development.

Regulation Must Evolve With the Product

A continuously improving medical device creates an obvious regulatory challenge. Traditional regulatory models are naturally suited to products whose capabilities are clearly defined at the time of approval. Software and AI can change after deployment, which means manufacturers need rigorous processes for determining when an update changes the clinical behaviour of the device and what validation is required before that update reaches users.

The solution cannot be to treat medical software like an ordinary consumer application. Updating a music streaming interface and updating an algorithm that contributes to a clinical decision involve fundamentally different levels of risk.

Medical device companies therefore need mature lifecycle governance. Software changes must be documented, tested and monitored. AI models need appropriate validation across relevant patient populations and clinical conditions. Performance after deployment needs to be observed so that unexpected behaviour can be identified and addressed.

Human oversight also remains critical. An AI-enabled device should make its role clear to clinicians. Healthcare professionals need to understand whether the system is providing a measurement, recommendation, prioritisation signal or another form of decision support.

This is another area where the implementation philosophy I have learned from Phaneesh Murthy becomes particularly relevant. Governance should not be treated as something added after innovation. In high-consequence environments, governance is part of the product architecture itself. A medical device company that can update software quickly but cannot validate those updates responsibly has not created a scalable innovation model.

Cybersecurity Becomes Part of Patient Safety

A medical device that is permanently disconnected from external systems presents one kind of security challenge. A connected device capable of receiving updates, exchanging information and participating in hospital networks presents another entirely.

Connectivity expands capability, but it also expands the potential attack surface.

Manufacturers therefore need to treat cybersecurity as part of the device lifecycle rather than a technical issue addressed primarily during development. Vulnerabilities may be discovered years after a device is installed. Operating environments change. New threats emerge. Hospitals connect equipment to increasingly complex digital infrastructure.

A software-defined medical device needs the ability to receive security updates throughout its supported life, and manufacturers need processes for identifying vulnerabilities and responding quickly.

This also creates a responsibility shared across the healthcare ecosystem. Device manufacturers, hospitals, technology providers and network operators all influence the security of connected clinical environments. A device can be well designed but still operate inside a poorly secured network, while strong hospital security cannot compensate for an unsupported device containing known vulnerabilities.

As Phaneesh Murthy puts it, “When a product becomes connected, responsibility does not end at the boundary of the product. You have joined an ecosystem, and your risk becomes part of everyone else’s risk.”

For medical devices, that interconnected responsibility is particularly important because cybersecurity can directly affect continuity of care and patient safety.

The Business Model Could Move Beyond the Initial Device Sale

If medical devices continue gaining capabilities after installation, commercial models may evolve as well. Manufacturers have traditionally generated revenue through equipment sales, maintenance, consumables and replacement cycles. Software introduces the possibility of additional services, subscriptions and capability-based offerings.

A healthcare provider might purchase the physical device while subscribing to advanced analytical features. Software modules could be introduced for specific clinical workflows. Fleet management platforms could help hospitals monitor equipment across facilities, while predictive maintenance services could reduce downtime.

This creates opportunities for recurring revenue, but it also changes customer expectations.

If healthcare providers are paying continuously for software capabilities, manufacturers need to demonstrate continuous value. Product teams must keep improving the platform, support needs become ongoing and software reliability becomes part of the commercial relationship.

During my mentorship under Phaneesh Murthy, I have learned that recurring revenue should never be confused with simply recurring billing. A sustainable platform business needs to create recurring value. The customer needs to understand why the product is becoming more useful over time.

Medical device companies that make that transition successfully could build significantly deeper relationships with healthcare organisations than companies whose primary interaction occurs during procurement and replacement.

The Medical Device Could Become a Healthcare Platform

The most important change may ultimately be conceptual.

A medical device has traditionally been understood as a specialised tool designed to perform a defined clinical function. That definition will remain appropriate for many products, but an increasing number of devices may become nodes inside much larger healthcare platforms.

The device generates information. Software interprets it. AI identifies relevant patterns. Hospital systems provide additional context. Clinicians make decisions, and the resulting outcomes can inform future improvements. The value therefore emerges not from the hardware alone but from the entire system surrounding it.

During my learning journey under Phaneesh Murthy, I have come to understand that platform transformations often begin when companies stop defining their product according to its physical boundaries. The customer is not ultimately buying hardware, software or AI independently. They are buying an outcome that requires all of those elements to work together.

For medical device companies, this means the next generation of competition may not be determined solely by who manufactures the most advanced piece of equipment. It may increasingly depend on who creates the most useful intelligence around that equipment, who integrates most effectively into clinical workflows and who can improve the product responsibly throughout its life.

A medical device will still need excellent hardware. It will still need reliability, clinical evidence and careful engineering. Those requirements do not become less important simply because software becomes more prominent.

But the idea that the product is finished when it leaves the factory may gradually disappear.

The medical device of the future could spend years inside a hospital and still become more capable than it was on the day it arrived.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy

For most of media history, taking a story global required an enormous amount of additional work. A film, television programme, advertisement, training video or digital series created for one market needed translation, subtitling, dubbing, editing and sometimes significant cultural adaptation before it could reach audiences elsewhere. The largest media companies could justify this investment for their most important properties, but the economics became considerably more difficult for smaller productions, niche audiences and regional content.

Artificial intelligence is beginning to change those economics. Translation models can work across languages with increasing sophistication. AI-assisted dubbing can accelerate voice production, while speech technologies can help adapt audio for different markets. Generative systems can support subtitling, localisation and versions of visual or written material designed for different cultural contexts. Work that previously required separate production processes for every market can increasingly begin with a common technological layer.

The most interesting consequence is not simply that Hollywood can distribute more efficiently around the world. It is that stories originating almost anywhere may have a more realistic path towards audiences almost everywhere. A regional production in India, a documentary from South America or an independent programme from Southeast Asia can potentially cross linguistic barriers that once restricted its commercial reach. During my learning journey under Phaneesh Murthy, I have increasingly understood that technology creates some of its greatest economic impact when it removes a constraint that businesses previously accepted as permanent. As Phaneesh Murthy puts it, “The most powerful technology shifts are often the ones that make an old limitation suddenly look unnecessary.” Language may be becoming one of those limitations.

Translation Is Only the First Layer of Localisation

Making content understandable in another language is not the same as making it feel native to another audience. Literal translation can preserve the meaning of individual words while losing humour, emotion, cultural references and the rhythm that gives dialogue its personality. Anyone who has watched poorly dubbed entertainment or encountered an awkwardly translated advertisement understands how quickly technically accurate language can damage the experience.

AI localisation therefore needs to extend beyond replacing one language with another. Systems increasingly need to understand context. A joke that works in one country may need to be rewritten rather than translated. A metaphor may have no equivalent meaning elsewhere. A phrase that feels warm and conversational in one language might sound unusually formal when converted directly into another. Even the length of translated dialogue can create challenges because the new sentence may need to fit the timing of the original performance.

This is where human linguistic and cultural expertise remains essential. AI can accelerate the first stages of translation, produce alternatives and help localisation teams work across much larger volumes of content. Human specialists can then evaluate whether the result actually communicates the intended emotion and cultural meaning.

During my mentorship under Phaneesh Murthy, I have learned that automation should be applied carefully when the quality of an outcome depends on context rather than simple correctness. Language is a perfect example. A machine may produce a technically correct translation, while an experienced local editor recognises that nobody in the target market would actually speak that way. The strongest localisation systems will therefore combine AI scale with human cultural judgment.

Dubbing Could Become Available to Far More Content

High-quality dubbing has historically been expensive. It requires translators, script adapters, voice actors, recording studios, directors, sound engineers and significant coordination. For major films and television programmes with large international audiences, the investment can be justified. For smaller productions, the economics often mean relying on subtitles or limiting distribution to fewer markets.

AI-assisted dubbing could dramatically change this equation.

Speech technologies can help generate translated audio, match timing more efficiently and provide production teams with rapid drafts before final versions are created. Voice performers could potentially work with AI tools to create approved versions across languages while retaining greater consistency in character and tone. Smaller media companies may therefore be able to consider dubbed distribution in markets that would previously have been commercially unrealistic.

The strategic opportunity extends beyond entertainment. Educational content, corporate training, product demonstrations, news formats and digital media can all become accessible to much wider audiences when localisation costs decline.

The key will be preserving performance. A voice is not simply a delivery mechanism for words. It contains emotion, personality, timing and identity. Media organisations that treat dubbing purely as a technical conversion risk producing content that is understandable but emotionally flat.

As Phaneesh Murthy says, “AI should reduce the cost of the process without reducing the value of the experience.”That distinction will matter enormously as AI-assisted dubbing becomes more common.

Regional Content Could Become Global Content

The economics of traditional international distribution have historically favoured content created in languages with large existing markets or productions backed by companies capable of investing heavily in localisation. AI could weaken that structural advantage.

India offers a particularly interesting example because the country’s media landscape already operates across numerous languages and highly distinct cultural markets. A programme created primarily for a Malayalam, Tamil, Telugu, Bengali or Marathi audience may have strong storytelling potential beyond its original linguistic market, but localisation and distribution add cost and complexity.

If AI reduces those barriers, regional content can travel more easily both within India and internationally.

This does not mean every piece of content will suddenly become globally successful. Cultural specificity can sometimes limit international appeal, while in other cases it becomes precisely what makes a story distinctive. The important change is that language itself becomes less of an automatic barrier to experimentation.

Media companies can test content in additional markets at lower cost, analyse audience response and invest further when genuine demand emerges. Smaller productions therefore gain access to a distribution strategy that was previously more accessible to large studios.

This reflects another idea I have taken from Phaneesh Murthy: when technology lowers the cost of experimentation, companies should not simply use it to do the same things more cheaply. They should increase the number of ideas they are willing to test. AI localisation could allow media companies to experiment with markets they would previously have ignored because the upfront economics were too uncertain.

Localisation Will Extend Beyond Language

A truly local experience involves far more than words. Visual references, music, pacing, humour, celebrity associations, cultural expectations and even the way stories are marketed can vary significantly between regions.

AI can help media companies understand these differences and create variations around them.

A trailer for the same programme might emphasise different characters or themes in different markets. Promotional artwork could highlight elements more likely to resonate with a particular audience. Descriptions, thumbnails and social content can be adapted while the underlying programme remains unchanged.

Advertising already operates this way to some extent, but generative AI makes the production of these variations significantly easier.

The risk is over-optimisation. If every creative decision is modified according to predicted local preferences, the original identity of the work can gradually disappear. Media companies therefore need to determine which elements belong to the core creative property and which can adapt around it.

During my learning journey under Phaneesh Murthy, I have come to appreciate the importance of separating a platform’s core from its configurable layers. The same principle can be applied creatively. The story, characters and artistic intent may remain fixed, while packaging, promotion and certain elements of localisation adapt according to market context.

That balance allows content to travel without asking it to become something entirely different every time it crosses a border.

AI Could Make Localisation Continuous Rather Than Sequential

Traditional global distribution often follows a sequence. Content succeeds in its original market, international potential is evaluated, localisation begins and additional markets receive the programme later. The process can create significant delays between the original release and broader availability.

AI can compress this timeline.

Translation and localisation can increasingly begin during production rather than after it. Scripts can be analysed for linguistic challenges before filming is complete. Draft subtitles can be generated quickly. Dubbing workflows can begin earlier, while promotional assets can be prepared for multiple markets simultaneously.

This makes global release strategies more accessible.

A production does not necessarily need to wait for success in one market before localisation work begins elsewhere. Media companies can prepare for simultaneous or near-simultaneous releases across a much broader group of regions when the commercial opportunity justifies it.

The strategic implication is important because audiences increasingly participate in global cultural conversations in real time. Social media can make a programme internationally visible within hours, and long delays between markets can reduce momentum or encourage unofficial distribution.

As Phaneesh Murthy puts it, “Technology creates strategic value when it removes waiting from a process where timing influences the outcome.” In entertainment, timing increasingly determines whether audiences participate in the same cultural moment or encounter content after the conversation has already moved elsewhere.

Human Local Experts May Become More Important, Not Less

At first glance, increasingly capable translation systems might appear to reduce the need for human localisation professionals. In practice, their work may shift towards the parts of localisation where judgment creates the greatest value.

Instead of spending the majority of their time producing first-pass translations, language experts can review AI-generated versions, refine cultural meaning, resolve ambiguity and protect the tone of the original work. Local creative teams can focus on understanding how a property should be positioned rather than manually adapting every individual asset.

This could make high-quality localisation more scalable.

A smaller team supported by intelligent tools may be able to supervise significantly more content while maintaining standards that fully automated systems would struggle to achieve. The role evolves from production towards direction and quality control.

This follows a pattern I have repeatedly observed through the enterprise perspective taught by Phaneesh Murthy. When AI handles more of the repetitive execution within a professional workflow, human expertise often moves upward. People spend less time producing the first version and more time deciding whether the version is appropriate.

In media localisation, cultural judgment becomes even more valuable precisely because the production of language becomes easier.

Rights, Consent and Creative Ownership Need to Travel With the Technology

AI localisation also introduces difficult questions around voice, likeness and creative ownership. If a performer’s voice can be replicated across languages, the media organisation needs clear agreements about how that capability can be used. If AI modifies a performance to match translated dialogue, artists should understand the scope of those modifications and the rights associated with them.

These questions cannot be treated as technical details.

A performer may be comfortable allowing an approved AI system to create foreign-language versions of their voice under specific conditions while objecting to unrestricted reuse elsewhere. Writers, directors and other creative professionals may similarly have legitimate concerns about how original work is modified during automated localisation.

Media companies therefore need governance frameworks that evolve alongside the technology. Consent, compensation, attribution and creative approval need to be considered as part of the production model rather than after disputes emerge.

Phaneesh Murthy has often emphasised that enterprise AI succeeds when governance is designed into the operating model rather than added after implementation. The same principle applies here. The ability to generate something does not automatically establish the right to generate it.

Companies that address these questions transparently may ultimately build stronger relationships with both creative professionals and audiences.

The Next Global Hit Could Begin Almost Anywhere

For most of media history, geography and language placed meaningful limits on distribution. Exceptional stories could cross those boundaries, but doing so required significant investment and a commercial decision that the potential audience justified the cost.

AI is reducing that friction.

A story can increasingly be translated, dubbed, packaged and marketed across multiple languages without requiring the same scale of incremental production that international expansion once demanded. Media companies can test additional markets more economically, while independent creators and regional studios gain access to capabilities previously concentrated among larger organisations.

During my mentorship under Phaneesh Murthy, I have learned that one of the most interesting consequences of technological change is what happens when distribution expands faster than production structures adapt. New participants gain access to markets, established players encounter competitors they previously never considered and audiences discover products that historically would never have reached them.

Media may be entering exactly that period.

The next globally successful programme does not necessarily need to begin in Los Angeles, London or another traditional media centre. It could originate in Chennai, Seoul, Lagos, Buenos Aires or a city that has never previously exported entertainment at significant scale. What matters increasingly is whether the story connects with people.

Language will still matter. Culture will still matter. Local expertise will still matter enormously. AI does not make the world culturally uniform, and it should not attempt to.

What it can do is make the distance between cultures easier to cross.

For media companies, that creates a remarkable possibility: one story can remain deeply rooted in the place where it was created while becoming understandable, accessible and emotionally meaningful to audiences in a hundred different markets.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy

Artificial intelligence is making it dramatically easier to produce creative material. Images can be generated in seconds, rough video edits can be assembled quickly, research can be condensed, scripts can be developed collaboratively with AI, audio can be processed automatically and content can be adapted for multiple channels with far less manual effort than before. For media companies, advertising businesses, studios and creative teams, this represents an extraordinary productivity opportunity. Work that once required days of repetitive production can increasingly be completed in hours.

But there is an interesting economic consequence to making content easier to produce. When the cost of producing something falls, the amount of it tends to increase. Audiences are therefore likely to encounter more articles, videos, advertisements, social content, entertainment and branded media than at any previous point in history. The competitive challenge for media companies may gradually shift from producing enough content to producing anything that deserves attention in an environment of almost unlimited supply.

This is why I believe the most important conversation about AI and creativity is not whether machines can create. They clearly can. The more important question is what happens to the value of human originality when competent creative production becomes widely accessible. During my learning journey under Phaneesh Murthy, I have increasingly understood that technological abundance often increases the value of whatever remains genuinely scarce. As Phaneesh Murthy puts it, “When technology makes execution available to everyone, differentiation moves somewhere else. The advantage shifts towards judgment, ideas and the ability to decide what is worth creating.”

For media companies, that shift could fundamentally change the economics of creativity.

AI Is Reducing the Cost of the Creative Process, Not the Importance of Creativity

Creative production contains far more operational work than audiences usually see. A finished piece of media may require research, transcription, editing, resizing, subtitling, colour correction, asset management, localisation and numerous rounds of revision. These activities are necessary, but they can consume enormous amounts of time that highly skilled creative professionals might otherwise spend developing ideas.

AI can accelerate many of these processes. Editors can use intelligent tools to organise footage and identify useful sequences. Writers can use AI to explore research or challenge early drafts. Designers can generate rough visual directions before developing the strongest concepts properly. Production teams can create transcripts, subtitles and variations much more efficiently. Media archives can become searchable through semantic understanding rather than relying exclusively on manually created metadata.

The immediate benefit is productivity, but the more interesting consequence is what creative teams do with the time that productivity releases. If a designer spends fewer hours manually creating repetitive variations, there is more time to develop the central idea. If an editor can locate relevant footage in minutes rather than hours, more attention can go towards storytelling. If a writer can accelerate background research, more energy can be spent developing a distinctive argument.

This is where the perspective I have learned from Phaneesh Murthy becomes useful. Automation should not be evaluated only according to how much labour it removes. It should also be evaluated according to what higher-value work becomes possible once routine effort is reduced. In creative industries, that higher-value work is often the part audiences actually care about.

The Volume of Average Content Is About to Become Enormous

AI dramatically lowers the barrier to competent production. A small company can now create visual material that previously required specialised resources. An individual creator can produce across formats that once required several people. Media organisations can generate far more variations of existing assets without proportionally increasing production teams.

This democratisation is valuable, but it also creates an inevitable increase in supply.

Consumers already live inside highly competitive attention environments. Social platforms contain endless feeds. Streaming services offer enormous libraries. Brands compete through email, search, video, podcasts and dozens of digital channels. As AI makes production easier, each of these environments can become even more crowded.

The result is a paradox. Content becomes easier to make while attention becomes harder to earn.

Media companies therefore cannot assume that greater production automatically produces greater engagement. Generating twenty variations of an idea does not make the underlying idea twenty times more interesting. Publishing more frequently does not guarantee that audiences will care more frequently.

As Phaneesh Murthy says, “AI can reduce the cost of creating an output. It cannot reduce the cost of earning someone’s attention.”

That distinction should become central to AI strategies within media organisations. Productivity matters, but productivity without differentiation can simply produce more material for audiences to ignore.

Human Taste May Become One of the Most Important Competitive Advantages

AI can generate possibilities at extraordinary speed. The harder question is which possibility should be chosen.

This is where taste becomes strategically important. Great creative professionals do not simply produce assets. They understand which ideas feel familiar, which feel fresh, what should be removed, when something is trying too hard and when a concept contains the potential to become culturally meaningful. These judgments are difficult because they involve context, intuition and an understanding of audiences that is not always visible in historical data.

As generative systems become capable of producing hundreds of reasonable options, the ability to select among them becomes increasingly valuable. A creative director may spend less time asking how to technically create something and more time determining which direction is actually worth developing. Editors may become more important because an abundance of material requires stronger selection. Brand leaders may need clearer points of view because AI can easily generate content that is polished but indistinguishable from everything surrounding it.

During my mentorship under Phaneesh Murthy, I have learned that intelligence inside an organisation does not eliminate the need for judgment. In many situations, it increases it. When technology gives a team ten possible answers instead of one, someone still needs to understand which answer aligns with the organisation’s strategy, customer and identity.

Creative industries will experience this intensely. AI can generate possibilities. Human taste determines which possibilities deserve to become real.

AI Should Become a Creative Partner, Not a Creative Autopilot

There is a meaningful difference between using AI inside the creative process and delegating the creative process to AI.

A writer might use AI to examine alternative structures, challenge assumptions or explore information before developing an argument. A filmmaker could use generative tools during storyboarding or previsualisation. A designer might rapidly explore multiple visual directions before selecting one for deeper development. A music producer could experiment with arrangements or sound variations that would otherwise require significantly more time.

In each case, AI expands the creative search space.

The danger appears when organisations optimise primarily for speed and volume. If AI is asked to continuously generate finished material based on previous successful patterns, creative output can begin converging towards what already exists. The system becomes excellent at reproducing familiar structures precisely because those structures are well represented in its training and organisational data.

Media companies therefore need creative workflows that deliberately preserve human intervention at the points where originality matters most. The machine can accelerate exploration, but people should define the intent, determine the standard and decide when something has become distinctive enough to publish.

As Phaneesh Murthy puts it, “The right role for AI is often to widen the set of choices available to an expert, not to make the expert irrelevant to the choice.”

For creative organisations, that principle could provide a useful boundary between productive augmentation and generic automation.

Personalisation Will Create Multiple Versions of the Same Creative Idea

AI also changes the economics of creative variation. Historically, producing separate versions of a campaign, programme or piece of content for different audiences could become expensive very quickly. Every additional variation required editing, design, copywriting, production and review.

Generative systems can reduce those marginal costs significantly.

A media company could adapt promotional material according to audience interests. A publisher could present different summaries or entry points into the same reporting. Entertainment platforms could create trailers that emphasise different aspects of the same programme depending on what is likely to interest a particular viewer. Brands could adapt creative executions across customer groups while preserving a consistent central idea.

The strategic challenge is ensuring that personalisation does not fragment the creative concept.

A strong campaign often succeeds because it has a recognisable point of view. If AI generates endless variations optimised independently for every segment, the brand can gradually lose coherence. Personalisation therefore needs boundaries. The core idea should remain consistent while execution adapts intelligently around it.

This is another area where the enterprise thinking I have learned from Phaneesh Murthy becomes particularly relevant. Scale and personalisation are not necessarily opposites, but they require a strong underlying system. The organisation needs to define what can change and what must remain consistent.

For media companies, the same principle applies to creative identity. AI can create hundreds of versions, but the organisation still needs to know what makes all of those versions unmistakably its own.

Media Archives Could Become New Sources of Creative Value

Large media organisations possess enormous archives accumulated across decades. Television networks hold footage, publishers possess photographs and articles, film studios maintain vast libraries of creative assets, while music companies manage recordings, artwork and related materials. Historically, much of this content has been difficult to reuse because finding relevant material across enormous archives requires extensive metadata and manual research.

AI can make these libraries significantly more accessible.

Multimodal systems can increasingly understand what appears inside video, audio, images and text. A producer could search for a concept, person, location, visual style or event and identify relevant material even when those characteristics were never manually tagged. Historical content can therefore become easier to discover, repackage and connect with new creative projects.

This creates both commercial and creative opportunities. Archival footage can support documentaries, social content and new programming. Publishers can connect current events with historical reporting more effectively. Entertainment companies can help audiences rediscover older material through contextual recommendations rather than relying solely on conventional categories.

But the same principles around rights, provenance and creative integrity remain essential. The fact that AI can find and transform an asset does not automatically establish that every transformation is appropriate or permitted.

During my learning journey under Phaneesh Murthy, I have come to appreciate that enterprise AI often creates unexpected value by making existing assets more usable. Organisations sometimes focus entirely on what AI can generate while overlooking what it can help them rediscover inside information and intellectual property they already own.

For media companies with decades of accumulated creative material, that may become an enormous opportunity.

The Economics of Creative Teams Will Change

As AI absorbs more production work, media organisations will inevitably reconsider how creative teams are structured. Some tasks will require fewer manual hours, while entirely new roles may emerge around AI-assisted production, model governance, creative systems and intellectual-property management.

The mistake would be to treat this purely as a headcount equation.

A company might technically be able to produce the same volume of content with fewer people. But if every competitor gains access to similar productivity tools, producing the same amount more cheaply may provide only a temporary advantage. The more interesting strategy may be to use the productivity gain to improve quality, experiment more aggressively and create formats that were previously too expensive or time-consuming to attempt.

Smaller teams could potentially execute larger creative ambitions. Independent creators may gain capabilities that once belonged only to major studios. Large organisations can experiment with niche audiences without requiring every concept to justify traditional production economics.

This could make the media environment considerably more competitive because production capability becomes more widely distributed.

Phaneesh Murthy has often emphasised that when technology changes the economics of an industry, leaders should resist using the old cost structure as the only framework for evaluating the opportunity. The more important question is what new business models become possible when the cost structure changes.

AI is rapidly changing the cost structure of creativity. The companies that benefit most may be those that use that change to create things they previously could not justify making.

In a World of Infinite Content, Point of View Becomes Scarce

Artificial intelligence will become extraordinarily good at creating content that is technically competent. Images will look professional. Videos will be polished. Copy will be grammatically correct. Music will sound convincing. The baseline quality of creative production may rise significantly.

That does not mean audiences will become easier to impress.

When competent execution becomes abundant, audiences may increasingly respond to the qualities that are harder to manufacture at scale: originality, authenticity, cultural understanding, emotional intelligence and a recognisable point of view.

During my mentorship under Phaneesh Murthy, I have learned to think about technological change in terms of shifting scarcity. When one capability becomes inexpensive and widely available, competitive advantage moves towards something else. In media, production itself may become progressively less scarce. The ability to create something genuinely worth paying attention to may become more scarce.

This is why the future of creative work should not be framed as humans competing with machines over who can produce more. Machines will almost certainly win any competition based purely on volume.

The more meaningful opportunity is collaboration. AI can accelerate research, experimentation, editing, localisation and production. Human creators can use that leverage to spend more time developing ideas, understanding audiences and making decisions that require taste.

Media companies may soon be able to create more content than at any point in history. The winners will not necessarily be the companies that take maximum advantage of that capacity.

They will be the ones that know when they have something worth creating.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy

For decades, travel companies have relied on a remarkably durable formula for loyalty. Customers fly, stay, spend and accumulate points. Those points can later be exchanged for flights, hotel nights, upgrades or other rewards, while status tiers provide additional benefits to the customers who spend most frequently. The model has been enormously successful because it gives travellers a tangible reason to consolidate their spending with particular airlines, hotels and travel brands.

But points are ultimately a representation of transactions, and genuine loyalty is more complicated than transaction frequency. A traveller may repeatedly choose a hotel because it is conveniently located rather than because they feel any meaningful attachment to the brand. Another customer may travel less frequently but actively prefer a particular airline, recommend it to others and choose it whenever circumstances allow. Traditional loyalty systems can struggle to understand the difference because they primarily measure what customers have already spent.

Artificial intelligence creates an opportunity to build a much richer understanding of the relationship. Travel companies can increasingly examine preferences, trip context, previous experiences, engagement patterns and customer behaviour alongside transactions. Instead of simply calculating how many points someone has earned, an intelligent loyalty system can begin understanding what the traveller actually values and determine which experiences are most likely to strengthen the relationship.

During my learning journey under Phaneesh Murthy, I have increasingly understood that customer loyalty should be treated as an outcome of accumulated experiences rather than a programme operated by the marketing department. As Phaneesh Murthy puts it, “A loyalty programme can reward a transaction, but loyalty itself is created when the customer repeatedly feels that choosing you was the right decision.” AI gives travel companies a new opportunity to understand what creates that feeling for each customer.

Traditional Loyalty Programmes Know What Customers Spent, Not Necessarily What They Value

Points programmes are excellent accounting systems. They know how many flights a traveller has taken, how many nights they have stayed, how much they have spent and which rewards they have redeemed. This information is useful, but it provides only a partial picture of the customer.

Two travellers with identical loyalty balances may value completely different things. One may care deeply about airport lounge access because they travel frequently for work. Another may value flexibility because family travel plans change regularly. A hotel guest might appreciate room upgrades, while someone else would gladly exchange that benefit for breakfast, late checkout or a more convenient cancellation policy.

Traditional loyalty programmes frequently solve this problem by creating large menus of benefits. AI creates the possibility of understanding which benefits actually matter to which customer.

The system can analyse previous choices, redemption behaviour, travel patterns and interactions to identify likely preferences. Instead of presenting every member with the same generic collection of rewards, the programme can increasingly emphasise benefits that are relevant to the individual’s travel behaviour.

This reflects an important lesson I have taken from Phaneesh Murthy about personalisation. Organisations often assume that personalisation means creating more variations of an offer. The more valuable form of personalisation is understanding which part of the value proposition actually matters to the customer and making that easier to access.

Recognition Could Become More Important Than Reward

Some of the most memorable loyalty experiences have very little to do with points.

A hotel recognising a returning guest’s room preference can create more emotional value than a small points bonus. An airline proactively helping a frequent traveller during a disruption can strengthen the relationship more than another promotional offer. A hospitality brand remembering that a family requires connecting rooms can remove friction from every future stay.

These moments communicate recognition rather than reward.

AI can help travel companies deliver this recognition consistently. Customer preferences and previous interactions can become available to employees and digital systems across the journey, allowing the organisation to respond with greater context.

The difference is subtle but important. A reward tells the customer, “You have spent enough to receive this benefit.” Recognition tells them, “We remember you.”

During my mentorship under Phaneesh Murthy, I have learned that enterprise customer experience becomes much more powerful when information travels with the customer. If the airline’s mobile application understands a traveller’s preferences but airport staff do not, the organisation has intelligence without continuity. If a hotel remembers the customer at one property but treats them as completely unknown at another, the brand relationship fragments.

AI can help create a shared memory across these interactions, allowing recognition to become part of the operating model rather than something dependent entirely on individual employees remembering individual customers.

AI Can Understand That Every Trip Has a Different Purpose

One of the limitations of traditional loyalty segmentation is that customers are often categorised according to relatively stable characteristics. They may be identified as business travellers, leisure travellers, high-value members or occasional guests. In reality, the same person can belong to several of these categories depending on the trip.

Someone who travels alone for business during the week may take family holidays several times a year. The preferences associated with those journeys can be completely different. A business traveller may prioritise speed, airport proximity, reliable Wi-Fi and flexible check-in. The same person travelling with family may care more about room configuration, activities, dining and convenience.

AI can help loyalty systems interpret this context.

Rather than assuming that historical preferences apply identically to every future trip, intelligent systems can combine customer history with information about the current journey. The destination, length of stay, travel companions, booking behaviour and timing can all help determine what the customer may value on this particular occasion.

This allows loyalty to become situational rather than purely demographic.

As Phaneesh Murthy says, “Understanding the customer is not enough if you ignore the context in which the customer is making the decision.” Travel is one of the clearest industries where context can completely change what the same person values from one week to the next.

Rewards Can Become More Personal Without Becoming More Complicated

One danger of increasingly sophisticated loyalty programmes is complexity. Airlines and hotels already operate systems involving points, tiers, qualifying spend, redemption categories, partner benefits and numerous conditions. Adding AI should not make customers responsible for understanding an even more complicated rewards architecture.

Ideally, AI should make the complexity less visible.

Instead of forcing customers to search through dozens of redemption possibilities, intelligent systems can identify options that provide meaningful value based on the customer’s circumstances. A traveller planning a long flight might receive an upgrade opportunity at a moment when it is particularly relevant. A family booking a holiday could receive a reward connected to dining or activities. A frequent hotel guest could be offered benefits based on the services they actually use.

The underlying loyalty system may become more sophisticated, but the customer experience should become simpler.

This is a recurring principle in the technology implementation perspective I have learned from Phaneesh Murthy. Complexity inside the enterprise should not automatically become complexity for the customer. The organisation can operate extremely sophisticated systems behind the scenes while presenting a clear, intuitive experience externally.

AI is particularly well suited to this role because it can evaluate thousands of possible offers and benefits while presenting the customer with only the few that are most likely to matter.

Loyalty Intelligence Could Become Particularly Powerful During Disruption

Travel does not always go according to plan. Flights are delayed, connections are missed, weather changes itineraries and hotel arrangements occasionally fail. These moments are frustrating for customers, but they are also some of the most important moments in the relationship between traveller and brand.

Traditional loyalty programmes often treat service recovery separately from loyalty. AI creates an opportunity to connect them.

If an airline knows that a high-engagement customer is likely to miss a connection, an intelligent system can identify alternative flights, evaluate available accommodation and present options before the traveller joins a queue for assistance. A hotel experiencing an operational problem could understand which guests are most affected and provide relevant alternatives based on their preferences.

The goal should not be to prioritise only customers with the highest status. Every traveller deserves appropriate support. The opportunity is to use customer intelligence to make the resolution more relevant and efficient.

A family travelling with young children may require a very different solution from a solo business traveller. AI can help service teams understand those differences quickly.

Phaneesh Murthy has often emphasised that customer relationships are tested most clearly when something goes wrong. As he puts it, “Loyalty is not built only by delivering the expected experience. It is often strengthened by how intelligently and fairly you respond when the expected experience becomes impossible.”

For travel companies, disruption management could therefore become one of the most powerful applications of loyalty intelligence.

The Loyalty Programme Can Become an Ecosystem

Travel rarely involves a single company. A journey may include airlines, hotels, ground transportation, restaurants, attractions, payment providers and numerous other services. Traditional loyalty programmes have recognised this through partnerships, allowing customers to earn and redeem points across participating brands.

AI could make these ecosystems significantly more intelligent.

Instead of simply allowing points to move between partners, a travel ecosystem can begin coordinating experiences. An airline may know when the customer is arriving. A hotel knows when the room will be available. Transportation services understand the journey between them. Activities and restaurants form another layer around the destination.

With appropriate permissions and carefully designed data-sharing arrangements, AI can help connect these elements into a more coherent travel experience.

The opportunity is not to create a system where every company knows everything about the traveller. It is to share enough relevant information to remove unnecessary friction.

During my learning journey under Phaneesh Murthy, I have come to appreciate that ecosystems become powerful when each participant contributes a capability that improves the overall customer outcome. The company that successfully orchestrates those capabilities can create value beyond the boundaries of its own product.

For travel loyalty programmes, this could mean evolving from reward ecosystems into experience ecosystems.

Loyalty Must Still Be Earned, Even When AI Understands the Customer Perfectly

There is an important limitation to everything AI can do in loyalty. Understanding the customer does not guarantee that the customer will remain loyal.

A hotel can have perfect information about a guest and still lose them because the service is poor. An airline can offer highly personalised rewards and still frustrate customers through unreliable operations. A travel platform can understand someone’s preferences in extraordinary detail while another company simply provides a better product.

AI cannot compensate indefinitely for weaknesses in the underlying experience.

This is another reason why loyalty should not be viewed solely as a marketing function. It is influenced by operations, service, pricing, product quality, customer support and the hundreds of interactions that form the travel journey.

Phaneesh Murthy puts it simply: “Data can help you understand why a customer stays, but it cannot substitute for giving them a reason to stay.”

The role of AI is therefore to help organisations understand which experiences strengthen the relationship and where unnecessary friction threatens it. The company still needs to deliver.

The Future of Loyalty Is a Relationship the Customer Can Feel

Points will not disappear. They remain useful, familiar and economically powerful mechanisms for encouraging repeat business. Status tiers will continue to matter, particularly in industries where frequent customers value recognition and predictable benefits.

But AI allows travel companies to build something around those mechanisms that is much more sophisticated.

A loyalty programme can begin understanding what individual customers value, recognising the context of different trips, remembering preferences across interactions and responding intelligently when journeys do not go according to plan. Rewards can become more relevant, service can become more continuous and recognition can travel with the customer across the brand.

During my mentorship under Phaneesh Murthy, I have learned to think of customer intelligence as useful only when the customer can eventually feel its impact. It does not matter how sophisticated the customer data platform is if every traveller still receives the same experience.

That may ultimately become the test for AI-powered loyalty.

The customer should not need to know that an algorithm analysed their travel history or that an intelligent system selected a particular benefit. They should simply notice that the airline seems to understand what matters during their journey, that the hotel remembers the details they previously shared and that resolving a problem requires less explanation than it did before.

For decades, loyalty programmes have been extraordinarily good at remembering how much customers spend. AI gives travel companies an opportunity to become much better at remembering why those customers chose them in the first place.

And in the long run, that may be worth considerably more than another collection of points.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy

The finest hospitality has always felt personal. A returning guest arrives at a hotel and discovers that the team remembers the room they prefer, the food they enjoyed during their previous visit or the small request they made the last time they stayed. None of these gestures is technologically extraordinary. What makes them powerful is that they communicate recognition. The guest does not feel like another reservation moving through a property management system. They feel remembered.

The challenge is that this kind of personalisation has historically depended heavily on individual employees, institutional memory and the ability of a hospitality organisation to manually record and share guest preferences. It works beautifully when an experienced hotel employee recognises a returning guest. It becomes considerably more difficult when a hospitality group serves millions of travellers across hundreds of properties, restaurants and digital channels.

Artificial intelligence creates the possibility of bringing the attentiveness associated with exceptional hospitality to a much larger scale. An AI concierge could understand a guest’s preferences, travel context, previous experiences and current needs, then use that information to make the stay feel increasingly personal. The objective is not to replace the warmth of human hospitality with a chatbot. It is to make sure the organisation remembers enough about the guest for every interaction to begin with better context.

During my learning journey under Phaneesh Murthy, I have increasingly understood that personalisation is most valuable when technology disappears into the experience rather than becoming the experience itself. As Phaneesh Murthy puts it, “The best technology does not make the customer admire the technology. It makes the customer feel that the organisation understood what they needed.” Hospitality may become one of the clearest examples of this principle.

The Guest Journey Begins Long Before Check-In

A hotel stay may officially begin when a guest checks in, but the customer’s experience begins much earlier. They search for destinations, compare properties, examine rooms, consider restaurants, research transportation and think about what they want to do after arriving. By the time the reservation is confirmed, the guest has already made several decisions that provide useful context about the trip.

AI can help hospitality companies understand this journey more holistically. A traveller booking a short business stay has very different requirements from a family arriving for a week-long holiday. A couple celebrating an anniversary may value different recommendations from someone attending a conference. Even when two guests book exactly the same room, the purpose surrounding that reservation can make their expectations completely different.

An intelligent hospitality platform can begin adapting communication and recommendations according to this context. Instead of sending every guest the same pre-arrival email listing every service available at the property, the hotel can highlight the information most relevant to the individual journey. Airport transfers may matter to one guest, restaurant reservations to another and family activities to someone else. The hotel does not need to overwhelm the guest with options when it can intelligently prioritise the ones most likely to be useful.

This reflects a principle I have repeatedly encountered while learning from Phaneesh Murthy: personalisation should reduce the customer’s decision burden rather than simply increase the number of targeted messages they receive. In hospitality, that distinction is particularly important because the guest is often trying to relax, explore or concentrate on the purpose of the trip. The technology should make the experience easier rather than demand more attention.

The AI Concierge Can Understand Preferences Without Making the Guest Repeat Them

One of the frustrations that persists across large service organisations is repetition. Customers repeatedly provide information the company technically already possesses. A hotel guest may request a particular room location during one stay and need to request it again during the next. Dietary preferences may be shared with the restaurant but remain invisible to other relevant parts of the property. A guest may repeatedly ask for late checkout despite having demonstrated the same preference across several visits.

AI can help convert these individual interactions into persistent customer understanding.

If a guest consistently chooses quieter rooms, frequently requests additional pillows and prefers breakfast early in the morning, those patterns can become useful context for future stays. The system does not need to automatically assume every historical preference remains permanent, but it can use them to make better suggestions or prepare staff with more relevant information.

This is where hospitality can move from recording transactions towards building memory.

The difference is significant. A transactional system knows that a customer stayed in room 412 three months ago. An intelligent system can understand that the customer requested a quiet room away from the elevator, ordered breakfast at 7 a.m. and asked the concierge for running routes near the property. Those details reveal much more about how the guest prefers to travel.

As Phaneesh Murthy says, “Customer data becomes valuable when the organisation can turn memory into better service. Remembering what happened without changing the next interaction is simply record keeping.”

For hospitality brands, the ability to remember intelligently could become one of the most important dimensions of customer loyalty.

The Concierge Can Move From Answering Questions to Anticipating Needs

Traditional concierge service is largely responsive. A guest asks where to eat, how to reach a destination or what activities are available, and the concierge provides recommendations. AI allows some of that interaction to become anticipatory without making it intrusive.

Consider a guest arriving late in the evening after a long flight. The hotel may know the arrival time, understand that the property’s main restaurant will soon close and recognise from previous stays that the guest usually orders dinner after checking in. An intelligent system could offer appropriate dining options before the guest needs to search for them.

A family staying during a period of expected rain could receive suggestions for indoor activities. A business traveller with an early departure might be offered breakfast or transportation arrangements at an appropriate time. A guest who previously booked spa treatments during leisure stays could be shown availability without receiving generic promotional messages throughout the trip.

The distinction between intelligent service and excessive communication is critical. Anticipation only feels valuable when the recommendation is relevant and appropriately timed.

During my mentorship under Phaneesh Murthy, I have learned that timing is one of the most underestimated components of customer experience. A useful recommendation delivered after the customer has solved the problem themselves creates little value. The same recommendation delivered shortly before the need emerges can feel remarkably thoughtful.

AI gives hospitality organisations the ability to improve that timing across thousands or millions of guest journeys.

Human Hospitality Could Become More Human

There is a tendency to assume that introducing AI into hospitality means reducing human interaction. That may happen in certain transactional processes, particularly where guests prefer speed and convenience, but it would be a mistake to assume that the ultimate objective is a hotel with fewer meaningful human interactions.

Hospitality is fundamentally emotional. Guests remember how people made them feel, particularly during important trips, celebrations or difficult situations. AI cannot replace the genuine warmth of an excellent hotel employee who notices that something is wrong, adapts to an unexpected request or creates a memorable moment through empathy and judgment.

What AI can do is give that employee better context.

Imagine a front-desk employee being informed that a returning guest previously experienced a service problem and that the issue was resolved. The employee can acknowledge the history appropriately rather than forcing the guest to explain it again. A concierge could see that a guest is celebrating an anniversary and tailor recommendations accordingly. Restaurant staff could understand relevant preferences before the guest arrives.

Technology becomes the memory layer behind human service.

As Phaneesh Murthy puts it, “AI should remove the informational burden from people so they can spend more of their attention on the human part of the interaction.”

In hospitality, this could be one of AI’s most valuable contributions. Employees spend less time searching systems for information and more time actually serving the guest.

Personalisation Should Travel With the Guest Across the Brand

Large hospitality groups often operate numerous properties, restaurants, loyalty programmes and experiences. Yet customers can still feel as though they are meeting the brand for the first time whenever they enter a new property.

AI creates an opportunity to build a more continuous relationship.

A guest’s relevant preferences could follow them across the hospitality ecosystem. Someone who consistently chooses particular room characteristics should not necessarily need to establish those preferences independently at every hotel. A traveller’s dining interests, wellness preferences or preferred style of service could help new properties understand how to serve them even when the guest has never stayed there before.

This does not mean creating an inflexible profile that determines every future experience. Travel itself often involves exploration, and people may want completely different experiences depending on the destination or purpose of the trip. The intelligent system needs to understand both continuity and context.

The same guest may behave differently during a business trip, family holiday and weekend getaway. AI should therefore treat preferences as useful signals rather than permanent rules.

This is another area where the enterprise thinking I have learned from Phaneesh Murthy becomes particularly relevant. Customer intelligence becomes more valuable when it can travel across organisational boundaries. If every property, department and digital channel maintains its own isolated understanding of the guest, the company may possess enormous amounts of data while still failing to deliver a coherent relationship.

The real opportunity is to create a shared intelligence layer that allows the entire hospitality brand to recognise the customer.

The AI Concierge Will Need Strong Boundaries Around Trust

Hospitality personalisation depends on information that can quickly become sensitive. Travel patterns, location, spending behaviour, dining preferences and personal requests can reveal significant details about a person’s life. The fact that AI can analyse this information does not mean every possible use of it will feel appropriate to the guest.

Trust therefore needs to become part of the architecture.

Hospitality companies should be clear about what information is being remembered and how it improves the experience. Guests should have meaningful control over preferences and communication. Systems should distinguish between information that genuinely improves service and information that simply exists somewhere inside the organisation.

The strongest personalisation often comes from relatively simple signals. Remembering that a guest prefers a particular type of room can be valuable without constructing an unnecessarily intrusive profile around every aspect of their behaviour.

Phaneesh Murthy has often emphasised that the sophistication of an AI system should not be measured by how much data it can collect. It should be measured by the quality of the outcome it can create with the information that is genuinely necessary.

Hospitality brands that respect this principle are more likely to make personalisation feel thoughtful rather than invasive.

The Future Concierge May Be Everywhere and Almost Invisible

The traditional concierge occupies a physical desk inside the hotel. The AI concierge will be less visible because it can operate across the entire journey. It can help before arrival, support the guest during the stay and maintain relevant context after departure. It can exist inside an application, support hotel employees behind the scenes and coordinate with restaurants, transportation, housekeeping and other services without requiring the guest to understand the systems involved.

The most successful version may therefore be one that guests barely notice.

During my learning journey under Phaneesh Murthy, I have come to appreciate that mature technology often becomes less visible as it becomes more useful. Customers should not need to think about which database contains their preferences or which AI model generated a recommendation. They should simply experience an organisation that appears unusually good at understanding what they need.

Hospitality has always aspired to create that feeling. Great hotels have achieved it for generations through exceptional people who remember guests, anticipate needs and understand the importance of small details.

AI does not invent that philosophy. It gives hospitality companies a new way to scale it.

The future of the AI concierge is therefore not a machine standing between the guest and the hotel. It is an intelligence layer working quietly behind the entire experience, helping every part of the organisation remember a little more, anticipate a little better and give its people more opportunities to deliver the kind of hospitality guests remember long after they leave.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy

When most people think about artificial intelligence, they imagine enormous computing environments somewhere far away. A user sends information to an AI system, powerful servers process it in a centralised data centre, and the result returns through the internet. This architecture has powered much of the recent AI revolution because centralised cloud infrastructure offers extraordinary computational capacity and makes sophisticated models available to businesses without requiring them to build their own supercomputing environments.

But the next phase of AI will not exist only inside the cloud. Intelligence is beginning to move closer to the physical places where information is created: factories, hospitals, vehicles, stores, warehouses, farms, homes and cities. Cameras, sensors, machines and connected devices increasingly need to interpret information and make decisions quickly, sometimes without waiting for data to travel hundreds or thousands of kilometres to a centralised computing environment and back again.

This shift towards edge AI could significantly change the role of telecommunications companies. Telecom networks already connect these physical environments to the digital world. With 5G, edge computing and increasingly software-defined infrastructure, those networks can potentially become part of the computing architecture itself. Telecom companies therefore have an opportunity to become more than providers of connectivity. They could become one of the infrastructure layers upon which distributed artificial intelligence operates.

During my learning journey under Phaneesh Murthy, I have increasingly understood that major technology transitions often create value for companies that occupy an important position between previously separate systems. As Phaneesh Murthy puts it, “When computing moves closer to where the business process happens, infrastructure stops being something in the background and becomes part of the business decision itself.” Edge AI could bring telecom directly into that position.

Not Every AI Decision Can Afford a Trip to the Cloud

Cloud computing remains extraordinarily powerful, and many AI workloads will continue to depend on centralised infrastructure. The question is not whether the edge replaces the cloud. The more important question is which decisions need to happen close to where data is generated and which can comfortably happen elsewhere.

A manufacturing facility provides a useful example. Cameras on a production line may inspect thousands of products for defects. Sending every high-resolution image to a distant cloud environment could create significant bandwidth requirements, while the manufacturing process may require a response almost immediately. Processing at least part of that information locally allows the system to identify defects quickly and transmit only the information that needs to travel elsewhere.

The same logic applies to many other environments. A connected vehicle may need to respond to information immediately. A retail security system may need to interpret video locally. A hospital may operate applications where latency, reliability and data handling are particularly important. Industrial machines may generate enormous quantities of sensor information even though only a small percentage of those signals require broader analysis.

Edge AI allows some intelligence to operate where the event is actually happening. The cloud remains available for model training, deeper analysis, coordination and storage, while local infrastructure handles decisions where speed, privacy or bandwidth make proximity valuable.

From my mentorship under Phaneesh Murthy, I have learned that technology architecture should ultimately follow the requirements of the business process. The question should not be whether cloud or edge is technically superior. It should be where a particular decision can be made most effectively.

5G Could Become an AI Infrastructure Network

The importance of 5G is often described through faster mobile internet, but its broader significance lies in the kinds of connected environments it can support. Higher capacity, lower latency and the ability to manage large numbers of connected devices make advanced mobile networks particularly relevant for industrial and enterprise applications.

When these capabilities are combined with edge computing, the telecommunications network can become part of an AI delivery architecture.

Imagine a large logistics hub where autonomous vehicles, cameras, scanners and warehouse systems are continuously exchanging information. Some decisions need to happen locally, while others require access to enterprise systems elsewhere. A 5G network can provide the connectivity layer, while nearby edge infrastructure provides computational capacity for AI workloads that cannot tolerate unnecessary delay.

The telecom provider can potentially manage both.

This is strategically important because enterprise customers generally do not want to assemble dozens of technologies simply to deploy a single intelligent application. If the network provider can combine connectivity, computing and orchestration into a coherent service, the complexity of deploying edge AI can fall considerably.

As Phaneesh Murthy says, “Enterprise technology creates the most value when sophisticated infrastructure becomes simple for the customer to consume.” Telecom providers have an opportunity to make distributed AI feel less like an infrastructure project and more like an accessible enterprise capability.

Factories Could Become One of the First Major Edge AI Environments

Manufacturing provides perhaps the clearest illustration of why edge AI matters. Modern factories generate enormous amounts of information through machines, sensors, cameras and operational systems. Many of the decisions associated with that information need to happen quickly and reliably within the facility itself.

Computer vision can inspect products as they move through production. Predictive systems can analyse machine behaviour for early indications of failure. Autonomous equipment can coordinate movement across factory floors. Safety systems can detect hazardous conditions, while intelligent scheduling platforms continuously respond to production changes.

These applications do not all need the same architecture. Some intelligence can operate directly on devices, some can run on computing infrastructure inside or near the facility, and other workloads can be sent to central cloud platforms for deeper analysis.

Telecom providers could help orchestrate this environment through private networks and edge computing platforms. Instead of simply connecting machines to the internet, they can provide the infrastructure through which industrial AI applications communicate and operate.

During my learning under Phaneesh Murthy, one recurring lesson has been that enterprise transformation becomes meaningful when technology moves from the IT layer into the operating core of the business. Edge AI represents exactly that shift. Artificial intelligence is no longer analysing factory performance after the fact. It begins participating directly in production, maintenance, quality and safety decisions.

For telecom companies, supporting those decisions creates a relationship with the enterprise that is far more strategic than selling connectivity alone.

Edge AI Can Also Solve a Data Movement Problem

The growth of connected devices creates an enormous amount of information, and sending all of that information continuously to centralised computing environments may not always be necessary or economically sensible.

Video is an obvious example. A network of high-resolution cameras can generate enormous quantities of data, even though only a small portion of that footage may contain events that require attention. Processing information locally allows AI to identify relevant events and transmit only what is necessary.

Industrial sensors create a similar challenge. A machine may generate thousands of readings while operating normally. The organisation may not need every individual measurement to travel across the network. Edge AI can analyse those signals locally and escalate patterns associated with unusual behaviour.

This creates a more selective information architecture.

Instead of asking networks to transport every piece of raw data, intelligence determines what deserves to move. That can reduce bandwidth requirements while also allowing applications to respond more quickly.

Phaneesh Murthy has often emphasised that the goal of enterprise intelligence should not be to centralise every possible piece of information. As he puts it, “Good architecture does not move data simply because it can. It moves the information required to make the next decision.”

That principle could become increasingly important as billions of connected devices generate more information than enterprises could reasonably transport and store indiscriminately.

The Edge Creates a New Security Challenge

Moving intelligence closer to devices also changes the cybersecurity landscape. A centralised cloud environment concentrates computing resources inside highly controlled infrastructure. Distributed AI creates thousands or potentially millions of additional points where software, devices and networks interact.

Each point becomes part of the security architecture.

Telecom companies already possess extensive experience securing networks and identifying unusual traffic patterns. AI could strengthen these capabilities by continuously analysing device behaviour and identifying activity that differs from expected patterns.

Consider an industrial environment containing thousands of connected sensors and machines. If one device suddenly begins communicating in an unusual way, that change may indicate malfunction, misconfiguration or malicious activity. An intelligent network can identify the anomaly and potentially isolate the device before the problem spreads.

Security therefore becomes deeply connected to network intelligence.

This could create another important enterprise service opportunity for telecom providers. Instead of offering connectivity and cybersecurity as separate products, they can potentially build security directly into the intelligent network layer.

During my mentorship under Phaneesh Murthy, I have learned that the strongest technology platforms frequently turn necessary operational capabilities into sources of additional customer value. Security is unavoidable in distributed AI. Telecom companies that make it intrinsic to the infrastructure may create a meaningful competitive advantage.

The Future Architecture Will Be Cloud, Edge and Device Together

Technology conversations sometimes frame architectural changes as replacements. Cloud replaced on-premise infrastructure. Edge computing will replace cloud. AI devices will eliminate the need for centralised processing. In practice, enterprise technology rarely evolves this neatly.

The more likely future is a layered architecture.

Some intelligence will operate directly on devices. Some will run within nearby edge infrastructure. Other workloads will continue to depend on massive centralised cloud environments. The challenge is determining where each workload belongs and moving information intelligently between those layers.

This creates an orchestration problem.

An application may need to operate at the edge during normal conditions but access cloud resources when deeper analysis is required. Models trained centrally may need to be distributed across thousands of edge environments. Updates must be managed consistently, security policies need to follow workloads, and performance must be monitored across the entire system.

Telecom companies could occupy a powerful position within this architecture because their networks connect devices, edge infrastructure and cloud platforms.

The opportunity is not necessarily to own every layer. It is to become the platform that helps enterprises manage the relationships between them.

This reflects an important strategic lesson I have taken from Phaneesh Murthy: companies do not always need to own the entire technology stack to capture significant value. Sometimes the most defensible position is to orchestrate the stack more effectively than anyone else.

Telecom Could Become the Physical Infrastructure of the AI Economy

Much of the current AI conversation focuses on models, software and computing chips. These technologies deserve attention because they represent extraordinary advances. But intelligence operating in the physical world requires another layer of infrastructure.

Factories need networks. Autonomous machines need reliable communication. Connected healthcare environments need secure data exchange. Smart cities need millions of devices to communicate. Retail stores, logistics networks and energy systems increasingly need intelligence to operate across physical locations.

Telecom infrastructure connects these environments.

During my learning journey under Phaneesh Murthy, I have come to see this as one of the most significant strategic opportunities available to the telecommunications industry. The network does not need to remain a passive pipe through which AI applications communicate. It can become part of the architecture that determines where intelligence operates, how quickly it responds and how securely it moves between systems.

The companies that capture this opportunity will need to think beyond traditional telecommunications products. They will need partnerships with cloud providers, hardware companies, AI platforms and enterprise software businesses. They will need new commercial models and deeper understanding of customer operations.

But the underlying advantage is already there.

Telecom companies spent decades building infrastructure that reaches almost everywhere people and businesses operate. As artificial intelligence moves beyond data centres and begins entering the physical economy, that infrastructure may become more strategically important than ever.

AI is leaving the data centre. The next question is who will build the roads it travels on.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy

For most of the modern telecommunications era, the industry’s fundamental product has been connectivity. Telecom companies built enormous networks, connected billions of people and businesses, and created the infrastructure that allowed the digital economy to emerge. The technologies evolved from voice to mobile data, broadband, fibre and 5G, but the commercial proposition remained largely recognisable. Customers paid telecom providers to connect devices, people and organisations to networks.

Artificial intelligence creates an opportunity for that proposition to expand considerably. Telecom companies possess something that many emerging technology businesses are still trying to build: infrastructure distributed across cities and countries, relationships with millions of customers, enormous volumes of network intelligence and deep experience operating mission-critical systems at scale. Those assets could allow telecom providers to move beyond selling connectivity and begin providing the infrastructure, intelligence and enterprise services required to operate increasingly AI-driven economies.

This does not mean connectivity becomes less important. Quite the opposite. Reliable connectivity becomes even more critical as AI applications move from centralised cloud environments into factories, vehicles, hospitals, stores, homes and public infrastructure. The strategic question is whether telecom companies remain primarily the networks carrying this intelligence or become active platforms enabling it.

During my learning journey under Phaneesh Murthy, I have increasingly understood that some of the largest transformation opportunities emerge when companies reconsider what their existing assets allow them to become. As Phaneesh Murthy puts it, “Transformation is not always about building a completely new capability. Sometimes it begins by recognising that the capabilities you already built for one market can create value in an entirely different one.” Telecom may be approaching precisely that moment.

The Network Is Becoming More Than a Transport Layer

A traditional telecommunications network primarily exists to move information. A smartphone sends a request, the network transports that request, and data returns to the device. The network’s success is judged through measures such as availability, coverage, latency and capacity. These remain essential, but modern telecom infrastructure is becoming increasingly software-defined, distributed and intelligent.

5G, edge computing, network APIs and increasingly sophisticated network management systems allow telecom infrastructure to participate more actively in digital applications. Instead of every piece of information travelling to a distant cloud environment for processing, some computation can occur closer to where data is generated. For applications that depend on very low latency or need to process large amounts of information locally, this can create significant advantages.

AI makes this architecture much more important. A factory using computer vision to inspect products may generate enormous quantities of visual information. An autonomous machine may need to make decisions within milliseconds. A healthcare application may have strong requirements around reliability, privacy and data handling. Sending every interaction across long distances to a centralised data centre may not always be the most effective model.

Telecom companies already operate infrastructure close to many of these environments. The opportunity is to turn that proximity into a platform for enterprise computing and intelligence. During my mentorship under Phaneesh Murthy, I have learned to look at infrastructure not simply in terms of what it currently does, but in terms of what new services become possible because the infrastructure already exists. That shift in perspective could become crucial for telecom strategy.

Enterprise Customers Will Need More Than Bandwidth

Businesses are rapidly experimenting with AI, connected devices, automation and increasingly distributed digital operations. As these systems become more important, enterprises will need reliable ways to connect, manage and secure them. Buying additional bandwidth solves only one part of that challenge.

Consider a manufacturer operating several factories. It may use connected sensors to monitor equipment, computer vision to inspect production, autonomous machines to move materials and AI systems to analyse operational performance. The enterprise needs connectivity, but it may also need edge computing capacity, device management, cybersecurity, application orchestration and guarantees around network performance.

Telecom providers are positioned close to many of these requirements.

Instead of selling the manufacturer a connectivity contract and leaving the rest of the architecture to other technology providers, a telecom company could increasingly provide a broader enterprise platform. Private networks, edge infrastructure, managed security, IoT management and AI services could become part of an integrated offering designed around the customer’s operational environment.

This changes the commercial conversation. The telecom provider is no longer simply discussing the amount of data an enterprise consumes. It begins discussing which business processes depend on the network and how those processes can become more intelligent.

As Phaneesh Murthy says, “The moment a technology provider begins understanding the customer’s operating problem rather than simply selling its technical capacity, the relationship changes from vendor to strategic partner.” For telecom companies searching for growth beyond increasingly commoditised connectivity, that transition could be extremely important.

Network Intelligence Can Become a Product in Its Own Right

Telecommunications networks continuously generate operational information. They understand traffic patterns, device behaviour, network congestion, mobility patterns and service performance at extraordinary scale. Much of this information is currently used internally to operate networks more efficiently and improve customer experience.

AI can make that intelligence significantly more useful.

Telecom companies can use machine learning to predict network demand, identify unusual behaviour, optimise capacity and anticipate equipment failures. As these capabilities mature, some of the underlying intelligence can potentially support enterprise customers as well, provided privacy, security and appropriate data-use boundaries are rigorously maintained.

An enterprise operating thousands of connected devices, for example, may want to understand whether abnormal communication patterns indicate a malfunction or security problem. A logistics company may need highly reliable connectivity across specific routes. A large event venue may require temporary network capacity that responds dynamically to demand. The telecom provider’s understanding of network conditions can become part of the service rather than remaining invisible infrastructure.

The important distinction is between selling raw data and creating useful intelligence. Customers do not necessarily need access to enormous quantities of network information. They need answers to operational questions.

This reflects a principle I have repeatedly encountered while learning from Phaneesh Murthy. More information does not automatically create more value. Intelligence becomes commercially meaningful when it is translated into a decision, recommendation or outcome that the customer can actually use.

Telecom Can Become the Orchestrator of Distributed AI

Enterprise AI is often discussed as if intelligence exists primarily inside enormous cloud data centres. In reality, many future AI applications will be distributed across a combination of cloud, edge and device environments. Determining where computation should occur will depend on latency, cost, privacy, reliability and the amount of data being processed.

Telecom companies occupy a strategically interesting position within this architecture because their networks connect all three layers.

An AI application running inside a retail store might perform some processing locally, send more complex tasks to nearby edge infrastructure and rely on centralised cloud systems for broader analytics. A manufacturing application could operate similarly, while connected vehicles may require yet another combination of local and remote intelligence.

The complexity lies in orchestrating these environments.

Telecom companies could increasingly provide platforms that determine where workloads should run based on application requirements. Instead of the enterprise manually designing separate connectivity, edge and cloud architectures, the telecom platform could help manage the distribution of intelligence across the network.

As Phaneesh Murthy puts it, “The most valuable platform is often the one that removes complexity the customer should never have needed to manage in the first place.”

If telecom providers can make distributed AI simpler for enterprises to deploy, monitor and scale, they move significantly higher in the technology value chain.

AI Could Create an Entirely New Generation of Managed Services

Telecom companies already provide managed services to enterprise customers, including network management, security and communications infrastructure. AI creates an opportunity to expand these services considerably because intelligent systems can continuously monitor complex environments and respond to changes faster than traditional operating models.

A telecom provider could offer intelligent cybersecurity services that identify unusual network behaviour across connected devices. It could provide predictive infrastructure management for distributed enterprise networks. AI systems could automatically adjust network configurations based on changing application requirements, while enterprise customers receive recommendations about capacity, performance and operational risk.

This is particularly relevant for mid-sized businesses that want to adopt sophisticated technologies but do not have enormous internal technology teams. Building a private 5G network, edge infrastructure, IoT platform, cybersecurity capability and AI operations function independently may be unrealistic. Purchasing those capabilities as a managed platform could be considerably more attractive.

During my learning journey under Phaneesh Murthy, I have come to appreciate that enterprise technology markets often expand when complexity is packaged into something easier to consume. Cloud computing did not merely create new infrastructure. It changed how companies could access infrastructure. Telecom has an opportunity to apply a similar principle to distributed AI and intelligent connectivity.

The commercial model therefore moves beyond selling network capacity towards selling business capability.

The Biggest Challenge May Be Organisational Rather Than Technological

Telecom companies possess many of the assets required to pursue this opportunity, but owning infrastructure does not automatically create a platform business. The skills, sales models and operating structures required to sell connectivity can be very different from those required to sell complex enterprise technology solutions.

A connectivity product can often be defined through coverage, speed, capacity and price. An AI-enabled enterprise platform requires understanding the customer’s workflow, integrating with existing systems and sometimes redesigning how a business process operates. Sales cycles become more consultative, implementation capabilities become more important and partnerships with cloud providers, software companies and hardware manufacturers may become essential.

Telecom companies therefore need to decide where they genuinely want to compete. Attempting to build every component internally may create unnecessary complexity. In many cases, the stronger strategy may be orchestration: combine network infrastructure with technology from a broader ecosystem and provide enterprises with a coherent operating environment.

This is another area where the enterprise thinking I have learned from Phaneesh Murthy becomes relevant. Strategy is not simply deciding which markets appear attractive. It requires understanding which assets create a genuine advantage and where partnerships can fill capabilities that do not need to be owned.

The strongest telecom AI platforms may therefore be ecosystems rather than closed technology stacks.

Connectivity Can Become the Foundation Rather Than the Final Product

The telecom industry does not need to abandon connectivity to move beyond it. Connectivity remains the foundation upon which every additional service depends. The strategic opportunity is to recognise that the network can support a much broader collection of capabilities than communication alone.

A telecom company that provides connectivity, distributed computing, security, device management and AI orchestration begins to occupy a very different position within the enterprise technology landscape. Instead of being one component within a company’s digital infrastructure, it can become a platform through which increasingly intelligent operations are delivered.

During my mentorship under Phaneesh Murthy, one lesson has consistently influenced how I think about technology businesses: companies should periodically reconsider the boundaries of the market they believe they operate in. Those boundaries are often defined by historical business models rather than by the capabilities the organisation actually possesses.

Telecom companies historically built networks to connect people and machines. AI may allow those same networks to become infrastructure for decisions, automation and distributed intelligence.

The next generation of telecom competition may therefore not be defined entirely by who offers the fastest connection or the largest data allowance. It may increasingly depend on which providers can make that connectivity useful for an economy in which intelligence is distributed across millions of devices, businesses and physical environments.

The network will still carry the data. But the larger opportunity may be helping customers decide what that data should enable.

This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.

www.phaneeshmurthy.com

#phaneeshmurthy #phaneesh #Murthy