Retail has always involved an element of prediction. Long before a product reaches a shelf or appears on an ecommerce platform, someone has to decide what customers are likely to want. Merchandisers choose categories, products, colours, sizes, price points and quantities months before actual demand becomes completely visible. Those decisions combine historical sales data with experience, cultural awareness, intuition and an understanding of how customers are changing.
When those decisions are right, the results can define an entire season. When they are wrong, retailers can find themselves with excess inventory in one category while customers search unsuccessfully for products that are unavailable in another.
Artificial intelligence is introducing a new layer of intelligence into this process. Rather than relying primarily on historical sales to understand future demand, retailers can increasingly analyse a much broader set of signals, including search behaviour, product interactions, local demand patterns, social conversations, weather, purchasing behaviour and emerging cultural trends. The result is not the elimination of the merchandiser. It is the creation of a far more informed one.
During my learning journey under Phaneesh Murthy, I have increasingly understood that some of the strongest applications of AI occur when technology strengthens a decision that already depends heavily on human expertise. As Phaneesh Murthy puts it, “AI becomes powerful when it gives an experienced person the ability to see patterns that experience alone could never process at scale.”
Retail merchandising is an ideal example of that principle because the future of the function will depend on combining human understanding of customers and culture with machine intelligence capable of recognising patterns across enormous quantities of information.
Retail Has Always Been a Forecasting Business
Every assortment represents a collection of bets about the future.
A fashion retailer decides which silhouettes will remain relevant six months from now. A grocery chain estimates how demand for particular products will change across seasons. An electronics retailer determines how much inventory to allocate across different cities. A home furnishings company attempts to anticipate which materials, colours and categories will attract customers in the coming year.
Traditionally, these decisions have been supported heavily by historical performance. If a category performed well last year, the retailer may allocate more space and inventory to it this year. Merchandisers supplement those numbers with supplier conversations, market research, competitor analysis and their own understanding of customers.
The challenge is that the pace of consumer change is accelerating. Trends can emerge and disappear within months or even weeks. Online conversations can suddenly influence demand for products that previously occupied relatively small categories. Local markets can behave differently from national averages, while customers increasingly expect retailers to respond quickly to changing preferences.
AI gives retailers the ability to continuously analyse these movements rather than waiting for them to appear clearly in historical sales reports.
Phaneesh Murthy often emphasises that enterprise intelligence becomes more valuable as the time between a signal appearing and an organisation responding to it decreases. For merchandising, this means identifying meaningful changes in customer behaviour while there is still enough time to act on them.
The Signals Behind Demand Are Becoming Much Richer
A purchase is one of the clearest signals a customer can provide, but it is also relatively late in the decision journey. Before purchasing, customers search, browse, compare, save products, read reviews, interact with content and abandon shopping carts. Across millions of interactions, these behaviours create patterns that can reveal changing preferences before they become visible through completed transactions.
AI allows retailers to analyse these signals together.
A sudden increase in searches for a particular material may indicate emerging demand. Customers repeatedly viewing products in a specific colour family may suggest that preferences are shifting. Growing interest in a category within one city may indicate a local opportunity that national sales numbers would obscure. External information such as weather patterns, events and cultural conversations can provide additional context.
The objective is not to respond to every signal. Retail generates enormous amounts of noise, and individual behaviours can be misleading. The value of AI lies in identifying which combinations of signals have historically preceded meaningful changes in demand.
As Phaneesh Murthy says, “Data tells you what happened. Intelligence begins when you understand which signals matter before the outcome becomes obvious.”
That transition from reporting demand to anticipating demand could become one of the most important competitive advantages in retail.
The Assortment Can Become Local Rather Than National
Large retailers have historically benefited from standardisation. Centralised buying creates purchasing power, simplifies operations and allows brands to maintain consistency across hundreds or thousands of locations.
Customers, however, are not standardised.
A product that performs exceptionally well in Chennai may have very different demand in Delhi. Even stores within the same city can serve customers with different incomes, lifestyles and preferences. Ecommerce adds another layer because digital demand can be analysed at an extremely granular geographic level.
AI gives retailers the ability to balance the efficiencies of centralised merchandising with the intelligence of local demand.
Instead of allocating identical assortments across broad regions, retailers can use predictive models to determine which products, sizes, quantities and price points are most appropriate for individual stores or clusters. The core brand remains consistent, but the assortment becomes increasingly responsive to the community surrounding each location.
During my learning under Phaneesh Murthy, one implementation idea has remained particularly relevant: scale and personalisation should not automatically be treated as opposites. Technology can allow large enterprises to maintain the operational advantages of scale while making increasingly local decisions.
For retailers, this could mean that a thousand-store network begins behaving less like one enormous standardised business and more like a thousand locally intelligent storefronts.
AI Can Help Retailers Buy Less and Sell More
One of the most important consequences of better merchandising intelligence is its impact on inventory.
Retailers constantly balance two competing risks. Buying too little means missing sales when products sell out. Buying too much ties up working capital and eventually leads to markdowns, discounting or unsold stock.
AI can help improve that balance by connecting demand forecasting more closely with assortment planning, allocation and replenishment.
If demand for a product begins accelerating faster than expected, intelligent systems can identify the change and recommend adjustments to replenishment or allocation. If interest begins weakening, purchasing decisions can be revised before excess inventory becomes a larger problem. Inventory can also be moved between locations based on predicted demand rather than waiting for individual stores to accumulate obvious surpluses and shortages.
The goal is not perfect forecasting. Consumer behaviour will always contain uncertainty. The advantage comes from continuously learning and adjusting as new information becomes available.
Phaneesh Murthy is of the belief that AI creates significant enterprise value when it converts large periodic decisions into smaller continuous decisions. Merchandising has traditionally involved major commitments made well before demand becomes fully visible. AI allows some of those decisions to become more adaptive as the market evolves.
The Merchandiser’s Role Will Become More Strategic
Whenever AI enters a function built around expertise, there is an understandable question about whether the technology will eventually replace the professional.
In merchandising, I believe the more interesting outcome is the opposite.
AI can analyse enormous amounts of behavioural and commercial information, but retail is not simply mathematics. Customers respond to culture, aspiration, identity, aesthetics and emotion. A product can become desirable because of a cultural moment that has no obvious historical precedent. A collection can succeed because a creative team introduces something customers did not previously know they wanted.
Human merchandisers understand these dimensions in ways that cannot simply be reduced to historical patterns.
AI therefore becomes a collaborator. It can show that interest in a particular category is increasing, identify which customer groups are driving the change and model potential demand across locations. The merchandiser can then combine that information with knowledge of the brand, cultural context and creative direction to determine how the business should respond.
As Phaneesh Murthy puts it, “The objective is not to automate expertise out of the organisation. It is to give expertise better information and more time to make the decisions that require judgment.”
That could make merchandising more strategic. Less time is spent assembling reports and manually comparing historical numbers, while more time can be dedicated to understanding customers, shaping assortments and identifying new opportunities.
Merchandising Intelligence Needs to Connect With the Rest of Retail
The value of AI merchandising becomes much greater when it does not operate as an isolated function.
If an AI system identifies rising demand but supply chain teams cannot respond, the insight creates limited value. If merchandising changes an assortment but marketing continues promoting products based on an outdated plan, the customer experience becomes disconnected. If ecommerce identifies an emerging product trend but physical stores cannot access that intelligence, the retailer loses an opportunity to act across channels.
The implementation challenge is therefore to connect merchandising intelligence with supply chain, pricing, marketing, ecommerce and store operations.
This is an area where the enterprise implementation thinking I have learned from Phaneesh Murthy becomes particularly important. Organisations often deploy AI within individual functions because those projects are easier to define and measure. The larger transformation happens when intelligence begins moving across functional boundaries.
Imagine an emerging demand signal being identified by ecommerce behaviour. Merchandising evaluates the opportunity, supply chain models inventory availability, stores receive updated allocation recommendations and marketing adjusts campaigns towards locations where interest is strongest. What previously required several teams and planning cycles becomes a coordinated enterprise response.
That is when AI stops being a merchandising tool and starts becoming part of an intelligent retail operating model.
The Future of Merchandising Will Combine Instinct With Intelligence
Retail will always need people who understand products and customers.
The best merchandisers develop a sense for what will resonate before the evidence becomes completely obvious. They understand how products fit together, how customers perceive value and how a collection or assortment communicates the identity of a brand. Artificial intelligence should not eliminate those capabilities. It should give them a significantly richer foundation.
During my mentorship under Phaneesh Murthy, I have learned to think about AI as a way of increasing the range of information that can participate in a decision. That idea is particularly powerful in merchandising because decisions that were once based on relatively limited historical information can now incorporate millions of behavioural, commercial and contextual signals.
The result could be a retailer that understands demand earlier, localises assortments more intelligently, responds faster to changing preferences and carries inventory that is more closely aligned with what customers actually want.
The AI merchandiser will therefore not be an algorithm sitting somewhere inside the retailer and independently choosing products.
It will be a new partnership between computational intelligence and human instinct, where AI helps retailers see the market more clearly and experienced merchandisers decide what to do with that understanding.
This blog is curated by young marketing professionals who are mentored by veteran Marketer, and industry leader, Phaneesh Murthy.
#phaneeshmurthy #phaneesh #Murthy