How Walmart Is Replacing Traditional Product Search With AI: A ChatGPT and Agentic Commerce Case Study

Executive Summary: Walmart Is Moving From Search Bars to AI Conversations

Online shopping has followed the same routine for years. You open a store’s website, type a few words in the search box, scroll through a long list of products, apply some filters, open a handful of product pages, compare a few options, and then finally buy something.

Walmart is now building a very different kind of shopping experience.

Today, a shopper has to know exactly what to type. Walmart wants to change that. In its new model, a shopper can simply explain what they need in plain language, and AI helps them find the right products.

The big step in this direction came in October 2025. Walmart announced a partnership with OpenAI to build AI-first shopping experiences. The early plan was to let Walmart customers and Sam’s Club members shop through ChatGPT, so that finding a product and buying it would sit much closer together.

The idea is simple, but it matters.

Instead of typing:

“wireless headphones noise cancelling under $100”

a shopper might say:

“I need comfortable wireless headphones for travelling, with good noise cancellation and a budget below $100.”

The second version tells the AI a lot more. It shares the purpose, the priorities and the price limit. An AI shopping system can use all of that to shrink the list of choices, so the shopper doesn’t have to dig through hundreds of listings by hand.

This is the point where agentic commerce starts to change ecommerce.

A traditional online store mostly reacts to single commands. The customer searches, filters, compares and decides everything on their own. In an agentic model, AI takes part in more of that work. It understands what the customer is trying to achieve, helps research the options, fine-tunes its suggestions, and connects the shopper more directly to the buying step.

By March 2026, this was no longer just an announcement. OpenAI said Walmart was launching an experience inside ChatGPT that could take people from finding a product in ChatGPT to a personalised Walmart space. That space supports account linking, loyalty features and Walmart payments.

The timing matters because Walmart’s digital business is already huge.

In Q4 FY2026, Walmart reported that:

• Global ecommerce net sales grew 24% year over year.
• Ecommerce made up 23% of Walmart’s total net sales.
• Total revenue for the quarter came to roughly $190.7 billion.
• Global membership fee revenue grew 15.1% year over year.

The growth continued into FY2027. In Q2 FY2027, Walmart reported 23% global ecommerce growth, and ecommerce made up 24% of total net sales.

These numbers make Walmart’s AI plan far more important than a normal experiment.

This is not a small retailer trying out a new website feature. Walmart is bringing conversational, AI-assisted shopping into one of the biggest retail systems in the world.

The larger point is that ecommerce may be entering a new phase.

For the last twenty years, retailers fought for visibility in search engines, marketplaces and retail search results. In the new AI shopping world, a different question is becoming just as important:

Can an AI system understand what a customer needs well enough to recommend the right product, and then connect that recommendation straight to a purchase?

Walmart’s work with ChatGPT suggests that finding, comparing and buying products may slowly become one single conversation.

The Problem With Traditional Ecommerce Product Search

Traditional ecommerce search works very well when the customer already knows what they want.

Say someone wants a particular TV model, laptop or pair of shoes. Typing the product name into the search box is probably the fastest way to get there.

Things get harder when the shopper has a need but doesn’t know the exact product yet.

The usual journey looks like this:

Search → Product List → Filters → Product Pages → Comparison → Checkout

On paper, that looks efficient. In real life, every step adds more work for the customer.

Search Engines Depend Heavily on Keywords

Most ecommerce search systems start with the exact words the shopper types.

That creates a basic problem: customers often don’t know the right product terms.

Someone may know they need a light laptop for university, video calls and a bit of photo editing. But they may have no idea which processor, how much RAM, or which product category to search for.

So the customer has to turn a real-life need into keywords the search engine can understand.

Conversational AI flips part of that process. The shopper doesn’t have to learn product language first. The AI reads the situation and then links that need to the right products.

Large Catalogues Create Too Many Choices

A store like Walmart sells in a huge number of categories.

That gives customers more choice, but more choice doesn’t always make a decision easier.

One simple search can bring back dozens or even hundreds of similar products. The shopper then has to compare:

• Prices
• Brands
• Ratings
• Product specifications
• Sizes
• Colours
• Delivery options
• Seller information
• Availability

The more alike the products look, the harder it is to pick one.

This is especially tough for people who are researching a category for the first time.

Filters Help, but Customers Still Need to Know What Matters

Filters were one of the best upgrades ever made to ecommerce search.

They let shoppers cut down a big catalogue by price, size, colour, rating or brand.

Still, filters leave most of the thinking to the shopper.

Someone who knows a product well can tell which filters matter. A first-time buyer usually can’t.

Take a person buying their first robot vacuum. They might care about pet hair, carpets, mapping, battery life and self-emptying. But they may not know which of these matters most in their home.

A conversation handles this in a different way. The shopper first explains their situation:

“I have two dogs, mostly hard floors and one carpeted room. I want something that can clean daily without much maintenance.”

Now the job is no longer matching keywords. It is understanding what the user needs.

Product Research Often Means Opening Many Tabs

OpenAI has pointed out that one weakness of normal online shopping is the constant jumping between tabs, repeated product lists and different sources before you can decide.

A typical shopper may:

1. Search on Google.
2. Open a retailer.
3. Read product reviews.
4. Open a second retailer.
5. Compare prices.
6. Watch a review video.
7. Go back to the first retailer.
8. Check delivery.
9. Finally decide.

Each extra step adds friction.

OpenAI’s newer shopping experience tries to pull more of the discovery process into one conversation. Users can describe what they need, refine it, browse products visually, and compare options side by side using details like price, features and reviews.

That doesn’t mean every shopper will stop visiting normal websites.

It does mean AI can cut down some of the research that happens before a customer feels ready to buy.

Comparison Is Still Mostly Manual

Traditional ecommerce sites are great at showing products.

They are not always good at explaining which product suits a specific situation.

A shopper may see:

• Product A: $79
• Product B: $94
• Product C: $109

But the real question is rarely “Which one costs less?”

It is usually “Which one gives me the features I actually need without making me pay for things I’ll never use?”

That is a reasoning problem, not a simple search problem.

Conversational shopping can help here because customers can ask follow-up questions.

For example:

“Which of these is better for travelling?”

Then:

“Remove anything over $100.”

Then:

“Which one has the best customer reviews for comfort?”

The search grows along with the conversation. It doesn’t start from zero every time the requirement changes.

Too Much Choice Can Slow Down a Purchase

A big selection is usually a plus for ecommerce, but it can also cause decision fatigue.

When shoppers keep comparing near-identical products, they may put off buying, or leave without choosing anything.

Conversational AI tries to lower that friction by helping customers shrink a large catalogue down to a small set of relevant options.

OpenAI said its updated shopping experience can use things like a shopper’s budget, preferences and limits to bring up suitable products, and then let the shopper refine the results through conversation.

So the real shift is not just:

Search bar → Chatbot

It is:

Keyword matching → Understanding intent

That difference is at the heart of Walmart’s move toward AI-assisted shopping.

Walmart’s Digital Scale Before the AI Shopping Shift

Walmart’s move into conversational commerce matters because the company already had a large, fast-growing ecommerce business before ChatGPT became part of the customer journey.

AI is being added on top of a strong digital commerce engine. It isn’t being used to build one from scratch.

Ecommerce Is Already a Major Part of Walmart’s Business

In Q4 FY2026, Walmart reported 24% year-over-year growth in global ecommerce net sales, and ecommerce made up 23% of total net sales. Total company revenue for the quarter was about $190.7 billion.

A quarter later, global ecommerce growth picked up to 26% in Q1 FY2027, and ecommerce still made up 23% of total net sales.

By Q2 FY2027:

• Global ecommerce sales were up 23%.
• Ecommerce made up 24% of Walmart’s total net sales.
• Walmart earned about $187.9 billion in quarterly revenue.

These figures show that digital commerce is no longer a side channel for Walmart.

By Q2 FY2027, roughly a quarter of the company’s net sales was already coming through ecommerce.

That gives AI-driven product discovery a much bigger possible impact than it would have at a retailer where online sales are still small.

Walmart U.S. Ecommerce Kept Growing

Walmart U.S. also showed strong digital momentum.

In Q2 FY2027, U.S. ecommerce sales rose 24% year over year.

This growth came from several parts of Walmart’s commerce system, not from one single channel.

Store-fulfilled delivery grew by about 43%. This shows that Walmart’s physical stores are increasingly working as part of its digital delivery network.

This link between online discovery and physical stores could be very valuable in AI commerce.

An AI recommendation only has commercial value if the product can actually be bought and delivered without trouble.

Walmart already has the setup to connect product discovery with local stock and delivery at a large scale.

Marketplace Growth Widens the Choice Available to AI

Walmart’s marketplace is another big piece of the picture.

In Q2 FY2027, Walmart reported that U.S. marketplace sales grew 52% year over year.

A growing marketplace means Walmart can offer many more products without having to hold all of that stock itself.

For AI-driven shopping, a wide catalogue is very useful. A conversational system does a better job when it can weigh many products across different:

• Brands
• Price levels
• Features
• Sellers
• Customer needs

But as the catalogue gets bigger, accurate product data becomes even more important.

AI needs dependable information about what a product is, what it costs, whether it’s in stock, and what makes it different from the alternatives.

This is one reason product feeds and structured commerce data are gaining importance in AI shopping. OpenAI’s Agentic Commerce Protocol, for example, lets participating merchants share product feeds and promotion details, so their products can show up more completely in ChatGPT’s discovery experience.

Delivery Speed Is Now Part of Walmart’s Digital Advantage

Finding a product is only half of shopping. The customer also cares about how fast it arrives.

Walmart’s store network gives it an unusual ecommerce edge, because thousands of physical stores can also work as local delivery points.

In Q2 FY2027, store-fulfilled delivery grew about 43%. Express deliveries completed in under three hours made up about 37% of store-fulfilled orders.

This changes what an AI recommendation is worth.

Picture a customer asking:

“What ingredients do I need for dinner tonight, and can I get them delivered before 6 p.m.?”

A useful shopping agent will eventually need to know more than the products themselves. It may also need details about:

• Local stock
• Delivery windows
• Location
• Pricing
• Substitutions
• Customer preferences

OpenAI has said the Agentic Commerce Protocol is meant to support wider AI-native commerce features over time, such as personalisation, local availability and estimated arrival times.

For a retailer with Walmart’s physical reach, that direction is very important.

Advertising Is Growing Alongside Ecommerce

Walmart’s ecommerce system is also becoming more closely tied to its advertising business.

In Q2 FY2027, Walmart said its global advertising business grew 38%, while Walmart Connect grew 43% excluding VIZIO.

This matters because modern retail platforms are no longer just places where sales happen.

They are turning into ecosystems that combine:

Discovery + Retail Media + Marketplace + Customer Data + Fulfilment + Loyalty

AI could eventually change how these pieces work together.

A conversational shopping experience may help Walmart understand intent earlier in the customer journey. Instead of only seeing what a customer clicks, the system may be able to understand what the person is trying to get done.

That could make the discovery layer more valuable for both shoppers and merchants.

Walmart+ Adds a Loyalty Layer

Membership is another important part of Walmart’s digital plan.

In Q2 FY2027, Walmart reported that global membership fee revenue rose 17% year over year. Walmart+ fee revenue grew at a double-digit rate, and the company reported record Q2 net additions for the membership programme.

The Walmart and ChatGPT link becomes more interesting when you look at it next to this membership system.

OpenAI said Walmart’s in-ChatGPT experience supports account linking, loyalty and Walmart payments.

So conversational shopping is not being built as only a product search tool.

It is being tied to the customer’s wider relationship with the retailer.

Walmart Is Adding AI to an Already Powerful Commerce Engine

Put together, the numbers show how big the opportunity is.

By Q2 FY2027, Walmart had:

• $187.9 billion in quarterly revenue.
• 23% global ecommerce sales growth.
• Ecommerce making up 24% of total net sales.
• 24% Walmart U.S. ecommerce growth.
• 52% U.S. marketplace growth.
• About 43% growth in store-fulfilled delivery.
• 38% global advertising growth.
• 43% growth in Walmart Connect, excluding VIZIO.

These are not the numbers of an ecommerce company still looking for product-market fit.

They describe a well-established global retail platform that is now trying out a new way for shoppers to reach products.

That is why the Walmart and ChatGPT case study is so relevant.

The biggest change may not be that Walmart is putting AI on its website.

The bigger change is that AI itself is becoming a front door to the Walmart shopping journey.

Walmart’s OpenAI Partnership: The Turning Point

Walmart’s partnership with OpenAI marked a real change in how the company thinks about online shopping.

In October 2025, Walmart announced it was working with OpenAI to create what it called AI-first shopping experiences. The plan was not just to put another chatbot on Walmart’s website. The bigger goal was to make shopping more conversational, more aware of context, and less dependent on traditional keyword searches.

The partnership covered both Walmart customers and Sam’s Club members.

According to Walmart, shoppers would eventually be able to find and buy products through ChatGPT using Instant Checkout, so that parts of the shopping journey could happen right inside an AI conversation.

This was a big break from how ecommerce had worked before.

For years, most online shopping was built on the same basic pattern:

Search bar → Product results → Filters → Product page → Cart → Checkout

Walmart’s AI-first vision opened up another possible path:

Describe a need → AI understands the request → Relevant products appear → Shopper refines options → Purchase

This isn’t only about interface design.

It changes the role of the shopping system itself.

Traditional ecommerce usually waits for a precise instruction from the customer. AI-powered commerce can read a broader goal and help the customer decide what to do next.

From Reactive Search to Conversational Commerce

Traditional search is mostly reactive.

A shopper types something, and the retailer sends back a list of matching products.

If the shopper changes their mind, they usually have to rewrite the query, change the filters or start over.

Conversational commerce works differently.

The shopper can keep building on the same request.

For example:

Customer: “I need a coffee machine for a small apartment.”

The AI might help narrow that down.

Customer: “I don’t want anything too large, and I want to spend less than $150.”

Then:

Customer: “I mostly drink cappuccinos, so I need something that can froth milk.”

Every message adds more context. The conversation itself becomes part of the discovery process.

This matters because people often don’t start shopping with an exact product name. They start with a problem they want to solve.

Walmart’s partnership with OpenAI is designed around this behaviour.

In its October 2025 announcement, Walmart said ecommerce was moving away from the familiar search bar and long product lists, toward experiences that are more personal, more contextual and more multimedia.

Instant Checkout Brings Discovery and Purchase Closer

Another key part of the partnership was Instant Checkout.

In the past, AI assistants could help people research products, but the actual buying usually happened somewhere else.

A shopper might ask an AI for a recommendation, get a few options, and then leave the chat to visit a retailer’s website.

Instant Checkout shrinks that gap.

For eligible products and merchants, ChatGPT can support checkout right inside the conversation. The customer can confirm things like the order, shipping details and payment without necessarily leaving the chat.

The merchant still handles the real order, payment and delivery through its own systems. ChatGPT simply acts as the interface that connects the customer with the merchant.

For Walmart, this makes it possible to join together several stages that used to be separate:

Inspiration → Discovery → Recommendation → Selection → Purchase

That is one of the foundations of agentic commerce.

Why Walmart Calls This Agentic Commerce

Walmart has described its AI shopping direction as a move from systems that simply react to instructions toward systems that can learn, plan and anticipate what customers need.

That is a helpful way to understand agentic commerce.

It doesn’t just mean adding AI-written text to an ecommerce site.

It means using AI as an active layer inside the buying journey.

A shopping agent may eventually help with things like:

• Understanding what the customer needs
• Finding the right product categories
• Comparing suitable products
• Keeping budget and preferences in mind
• Refining choices through follow-up questions
• Connecting the customer to checkout
• Supporting repeat or routine purchases

This creates a more goal-focused experience.

Instead of asking, “What keyword did the customer type?” the system can start asking, “What is this customer trying to achieve?”

That is a much bigger change.

Walmart and Sam’s Club Bring Scale to the Experiment

The partnership is also significant because it includes both Walmart and Sam’s Club.

These two businesses serve different shopping habits, from everyday household buying to bulk purchases and membership-based shopping.

That gives conversational AI a wide range of possible uses.

One customer might ask:

“Help me plan five dinners for a family of four and build the shopping list.”

Another might ask:

“What household essentials should I buy in bulk to save money this month?”

A traditional search engine would need several separate searches for each of these.

A conversational system can potentially handle each one as a single connected task.

That makes Walmart’s AI strategy broader than plain product search.

It comes closer to real shopping help.

Personalisation Could Become a Major Competitive Advantage

Personalisation is another important part of the Walmart and OpenAI plan.

Traditional ecommerce personalisation usually relies on signals such as:

• Past purchases
• Browsing history
• Products viewed
• Location
• Saved preferences
• Membership activity

Conversational AI adds one more signal: what the customer is saying right now.

A customer can directly explain their budget, situation, priorities and limits.

That information can help create a much more specific shopping journey.

In March 2026, OpenAI announced that Walmart was launching an in-ChatGPT shopping experience that could take users from product discovery into a Walmart environment with features like account linking, loyalty and Walmart payments.

This is an important step.

It shows that the Walmart–OpenAI relationship is moving past basic product suggestions and toward a more connected commerce experience.

So the turning point is not just that Walmart teamed up with an AI company.

It is that Walmart is starting to treat conversation itself as a commerce interface.

From “Search for a Product” to “Tell AI What You Need”

The biggest difference between normal ecommerce search and AI-driven product discovery is easy to see with a simple example.

Traditional Search

“best running shoes women size 7”

This query hands the search engine a few keywords:

• Running shoes
• Women
• Size 7

The system then tries to match those words with products in its catalogue.

Now look at a conversational request.

Conversational Search

“I need comfortable running shoes for three short runs per week, mostly on pavement, under $100.”

This request carries far more information.

It tells the AI:

• What the shopper wants
• How often the product will be used
• Where it will be used
• What matters most
• The shopper’s budget

That extra context can completely change how products are found.

The AI is no longer just matching words against product titles.

It is trying to understand why the person is buying.

Intent Becomes More Important Than Exact Keywords

Traditional search often depends on the shopper picking the right words.

AI shopping can focus more on intent.

Think about someone searching for:

“small desk”

That could mean many things. The shopper may want:

• A desk for a small bedroom
• A desk for a child
• A foldable desk
• A desk for a laptop
• A standing desk
• A desk under a certain price

A conversational request clears up the confusion:

“I need a desk for a small bedroom. It only needs to hold a laptop and monitor, and I have about four feet of wall space.”

Now the AI understands a lot more than the word “desk.”

It understands the customer’s problem.

Context Helps Narrow Large Product Catalogues

Context is especially valuable for big retailers like Walmart.

When a store has thousands of products in one category, showing all of them isn’t very helpful.

AI can shrink that list using details from the conversation.

For example:

“I need a suitcase for a five-day business trip. I usually travel by air and want something that can fit in the overhead compartment.”

The system can pick up several pieces of context:

• Use case: Business travel
• Trip length: Five days
• Transport: Air travel
• Preference: Carry-on size

With a normal search, the user would have to turn all of these needs into filters themselves.

Conversational discovery lets customers say them naturally.

Budget Can Become Part of the Conversation

Price filters have been around in ecommerce for years.

But conversational AI can treat budget as part of a bigger decision, not just as a single filter.

Suppose a shopper says:

“I have $500 for a laptop, but I can spend slightly more if the battery life and performance are significantly better.”

This isn’t a simple maximum-price filter.

The customer is describing a trade-off.

An AI system can potentially weigh the extra cost against the extra value.

This kind of reasoning is very useful for purchases where the cheapest product isn’t always the best one.

Use Case Can Matter More Than Product Category

One of the strongest benefits of conversational shopping is that people can describe what they plan to do.

Compare:

“I need a blender.”

with:

“I make smoothies every morning with frozen fruit and ice, and I want something easy to clean.”

The second request contains buying criteria without using any technical specifications.

The AI can turn the customer’s use case into the right product attributes.

For example:

• Motor power
• Ice-crushing ability
• Container size
• Cleaning features
• Durability

This means shoppers don’t have to become product experts before they buy.

Previous Messages Can Refine the Search

Conversation also lets discovery develop naturally.

A customer might start with:

“Help me find a smartwatch.”

Then add:

“Battery life matters more than having lots of apps.”

Then:

“I also want something suitable for swimming.”

Then:

“Keep it below $250.”

Each message sharpens the AI’s understanding of the request.

The shopper doesn’t have to start the search again.

OpenAI’s shopping research feature is built to work this way. It can ask follow-up questions about things like preferred brands, size, performance, comfort, style or price, and users can keep adjusting their limits during the research.

Product Comparison Becomes Conversational

Comparisons on traditional ecommerce sites are usually fixed.

A comparison table may show:

Product	Price	Rating	Feature
Product A	$79	4.4	Feature X
Product B	$99	4.6	Feature Y
Product C	$119	4.5	Feature Z

But customers often want answers to more specific questions.

For example:

“Which one is easiest to use?”

or:

“Which one gives me the best balance between price and battery life?”

or:

“Remove the products that are too heavy for travel.”

Conversational discovery lets the comparison keep adjusting around what the shopper really cares about.

OpenAI says its shopping research tools can compare products side by side using details like price, features and reviews, while letting users remove products or change their needs along the way.

Why This Matters for AEO, GEO and AI Visibility

This shift has an important meaning for marketing.

Traditional SEO asks:

Does our product rank for the right keyword?

AI-driven discovery adds another question:

Can an AI system understand when our product fits a customer’s situation?

That is a much wider challenge.

A product may need to be understood not just through its title, but also through:

• Its purpose
• Features
• Benefits
• Price
• Use cases
• Audience
• Compatibility
• Reviews
• Availability
• Comparisons with alternatives

This is where ideas like Answer Engine Optimisation (AEO), Generative Engine Optimisation (GEO) and AI Visibility become more relevant.

Brands are no longer writing only for people who type exact keywords into Google.

They also need to prepare for people asking an AI assistant:

“What should I buy?”

So the marketing challenge starts to move from simply being searchable to being understandable and recommendable.

How Product Discovery Inside ChatGPT Actually Works

ChatGPT shopping is different from a traditional ecommerce search engine.

It doesn’t start with a fixed catalogue page. It starts with the user’s conversation.

When ChatGPT notices shopping intent, it can show products with images, product details and links to merchants. For some eligible products and merchants, checkout options may also be available right inside ChatGPT.

The process can be understood in several stages.

Step 1: The Shopper Expresses an Intent

Everything starts with what the customer wants.

The request may be very specific:

“Find me a 55-inch 4K television under $600.”

Or it may be much more open:

“I need a good TV for a bright living room, mostly for sports and streaming.”

The second request has fewer typical ecommerce keywords but far more useful context.

ChatGPT can use the request and the conversation around it to understand what the shopper is trying to do.

OpenAI says product results are chosen based on how relevant they are to the user’s intent and context.

Step 2: The Conversation Adds More Detail

If the first request is broad, the system can ask for more information.

For example:

AI: “What matters more to you: picture quality, price or gaming performance?”

The shopper might reply:

“Picture quality first, but I don’t want to spend more than $700.”

That answer helps narrow down the products.

This is one of the biggest differences between traditional search and conversational discovery.

A normal search engine expects the shopper to rewrite the query.

A conversational system can ask for clarification.

Step 3: ChatGPT Searches Across Product Information

Once the customer’s intent is clearer, ChatGPT can pull from several product-information sources.

According to OpenAI, shopping research may use:

• Merchant product data supplied through the Agentic Commerce Protocol
• Publicly available product information
• Other relevant retail sources

This means product discovery isn’t necessarily based on one fixed database.

The system can gather information from different commerce sources while it looks for suitable products.

Step 4: Relevant Products Are Matched to the Request

ChatGPT then tries to find products that fit the shopper’s needs.

Imagine the user says:

“I need a lightweight laptop under $900 with strong battery life for university and travelling.”

The system may judge products on things like:

• Price
• Weight
• Battery performance
• Portability
• Product specifications
• Reviews
• Availability
• Fit with the stated use case

The goal isn’t to return products that simply contain the phrase “lightweight laptop.”

It is to find products that match the whole request.

OpenAI states that product results are not picked because of advertising partnerships. Shopping results are generated separately, based on relevance.

Step 5: Products Can Be Presented Visually

Product discovery inside ChatGPT is becoming more visual.

Instead of getting only a paragraph of suggestions, users can see products along with details such as:

• Images
• Product names
• Prices
• Key features
• Reviews
• Merchant links

OpenAI expanded its shopping experience in March 2026 to offer richer visual browsing and comparison inside the conversation.

This makes shopping feel closer to a normal ecommerce site while keeping the conversation layer.

Step 6: Products Can Be Compared Side by Side

Comparison is one of the most useful parts of the system.

Instead of opening lots of browser tabs, the shopper can compare products within the same research session.

ChatGPT can bring up information like:

Product A

• Lower price
• Good reviews
• Shorter battery life

Product B

• Higher price
• Better battery life
• Lighter design

Product C

• Mid-range price
• Stronger performance
• Heavier body

The user can then ask:

“Which one is best for travelling?”

or:

“Show me only the options below $800.”

or:

“Remove anything with poor battery life.”

The comparison becomes interactive instead of fixed.

OpenAI says users can adjust shopping results while the research is still running, including removing products, asking for alternatives or changing their limits.

Step 7: Product Information Can Be Updated From Merchant Sources

Commerce data changes fast.

A product that costs $99 today might cost $119 tomorrow.

Stock can change within hours.

That makes fresh product information extremely important.

OpenAI’s Agentic Commerce Protocol lets merchants share product feeds and promotions, so product details show up more accurately and completely inside ChatGPT.

OpenAI also allows merchants to give direct product feeds to keep information about their products fresher.

This creates an important link between traditional ecommerce systems and AI discovery.

Product feeds have long mattered for channels like shopping ads and marketplaces.

They may now matter for AI-driven product discovery too.

Step 8: Merchant Integrations Connect AI Discovery With Commerce

The last stage is where AI discovery starts turning into commerce.

OpenAI’s Agentic Commerce Protocol works as a connection layer between merchants and ChatGPT.

Merchants can provide product information from their existing systems, and ChatGPT can use that information during product discovery.

OpenAI has said ACP can also support wider commerce experiences over time, including areas such as:

• Personalisation
• Local product availability
• Estimated arrival times

Walmart’s integration shows how this can work at scale.

Instead of ChatGPT just recommending Walmart products and sending customers away, Walmart’s in-ChatGPT experience can take the shopper from discovery into a personalised Walmart environment linked to account, loyalty and payment features.

The New Product Discovery Flow

The traditional journey might look like this:

Keyword → Search Results → Filters → Product Pages → Comparison → Retailer → Checkout

The new AI journey can look more like this:

Need → Conversation → Understanding Intent → Product Matching → Comparison → Refinement → Merchant → Purchase

On the surface the difference looks small. Strategically, it is big.

In the traditional model, the shopper does most of the work needed to find the right product.

In the conversational model, AI does a bigger share of the discovery and comparison work.

For retailers and ecommerce brands, this creates a new visibility challenge.

It may no longer be enough for a product to show up when someone searches its exact name.

The product increasingly needs to be understood when someone describes the problem it solves.

That is one of the clearest reasons Walmart’s move toward ChatGPT and agentic commerce deserves attention well beyond Walmart itself.

Walmart’s In-ChatGPT Shopping Experience Explained

By early 2026, Walmart’s work with OpenAI had gone beyond simply helping people find products in an AI chat.

On March 24, 2026, OpenAI announced a richer shopping experience inside ChatGPT and named Walmart as one of the companies bringing more of the retail journey into the conversation. The Walmart experience was designed to link product discovery in ChatGPT with a personalised Walmart shopping environment.

This matters because it brings several stages of ecommerce much closer together.

The new journey looks like this:

Customer Need → ChatGPT Conversation → Product Discovery → Walmart Experience → Account/Loyalty/Payment → Purchase

Each stage removes some of the friction that normally sits between finding a product and buying it.

Step 1: The Customer Starts With a Need

The journey doesn’t have to begin with a product name.

A shopper might say:

“I need a birthday gift for a 10-year-old who likes science, and I want to spend less than $50.”

In a normal ecommerce experience, the customer might have to search many terms:

• Science gifts
• STEM toys
• Gifts for 10-year-olds
• Toys under $50
• Educational toys

With conversational shopping, all of these conditions fit into one request.

The AI can use the customer’s goal, budget and the person they’re buying for to start narrowing the search.

This is a basic change, because the starting point is no longer just a keyword.

It is the customer’s intention.

Step 2: ChatGPT Helps Refine the Request

The shopper can carry on the conversation without starting a new search.

For example:

“Nothing too complicated.”

Then:

“It should be something they can use more than once.”

Then:

“Show me options with strong customer reviews.”

Each message adds another layer of context.

ChatGPT can use these details to fine-tune the products it shows.

OpenAI’s March 2026 shopping update was built around this kind of interaction. Users can describe what they want, refine their needs through conversation, and compare relevant options in the same experience.

This makes shopping feel less like moving through a database and more like getting help from a knowledgeable shop assistant.

Step 3: Product Discovery Happens Inside the Conversation

Once ChatGPT understands the request, it can show relevant products visually.

Instead of showing only links, the shopping experience can include details such as:

• Product images
• Product names
• Prices
• Features
• Availability
• Reviews
• Merchant information

Customers can also compare products side by side and keep narrowing their choices.

OpenAI said its March 2026 update brought richer visual browsing, side-by-side comparison, and more detailed, up-to-date product information right inside ChatGPT.

For shoppers, this reduces how many different tabs they need to open before deciding.

Step 4: The Shopper Can Move Into a Walmart Experience

This is where Walmart’s integration gets especially interesting.

Instead of ChatGPT just recommending a Walmart product and ending there, the customer can move from AI discovery into a personalised Walmart experience.

This builds a bridge between:

AI product discovery

and

Walmart’s commerce infrastructure

That difference matters.

ChatGPT helps understand and narrow the customer’s need, while Walmart provides the retail environment where the purchase can continue.

The experience therefore combines two different strengths:

• OpenAI: Conversation, understanding intent and product discovery
• Walmart: Catalogue, customer relationships, inventory, delivery, payments and loyalty

When these systems work together, AI becomes more than a recommendation tool.

It becomes a doorway into an established retail ecosystem.

Step 5: Account and Loyalty Can Join the Shopping Journey

Walmart’s in-ChatGPT experience is also designed to connect with customer accounts and loyalty features.

This matters because personalisation is more useful when an AI shopping experience can connect with an existing retailer relationship.

For example, a connected customer journey could potentially take into account things like:

• Existing Walmart account
• Loyalty benefits
• Membership status
• Saved preferences
• Previous interactions
• Relevant payment options

The exact experience will depend on what Walmart and OpenAI make available, but the direction is clear.

The shopping conversation is being linked to the customer’s wider retail profile instead of living as a separate research tool.

Step 6: Walmart Payments Bring the Journey Closer to Purchase

Payment is one of the last hurdles between a recommendation and a sale.

If a customer finds the right product in ChatGPT but then has to leave, search again, sign in somewhere else and type in details again, some of the convenience is lost.

The Walmart experience announced by OpenAI was designed to support Walmart payments as part of the journey.

This moves the customer closer to one continuous experience:

Ask → Discover → Compare → Select → Pay

The fewer unnecessary jumps there are between these stages, the easier buying can become.

Why This Is More Than a Chatbot Integration

It would be easy to call Walmart’s ChatGPT experience a chatbot partnership.

That would miss the bigger point.

The real development is the joining of AI discovery with retail infrastructure.

ChatGPT helps answer:

“What should I buy?”

Walmart helps answer:

“Can I actually buy it, get it quickly, and use my existing Walmart relationship to finish the purchase?”

Bringing these two questions together is one of the clearest examples of how conversational AI is starting to reshape ecommerce.

The new model isn’t simply:

AI recommends a product.

It is more and more:

AI understands the need, helps narrow the options and connects the customer directly to a commerce experience.

What Is Agentic Commerce, and Why Does It Matter?

The term agentic commerce can sound technical, but the basic idea is fairly simple.

Traditional ecommerce asks the shopper to do most of the actions.

The customer:

• Searches
• Filters
• Opens products
• Compares options
• Picks an item
• Goes to checkout
• Completes the purchase

The website gives tools, but the customer drives almost every single step.

Agentic commerce brings in AI systems that can help with more of this journey, based on the customer’s goal.

Put simply:

Traditional ecommerce waits for the shopper to act.

Agentic commerce lets AI help the shopper work toward a result.

From Commands to Goals

Traditional search works well with commands.

For example:

“Search for wireless earbuds.”

Agentic commerce becomes more useful when the customer states a goal:

“I need wireless earbuds for commuting. Noise cancellation is important, but I don’t want to spend more than $120.”

Now the AI has a task to work through.

It can potentially help work out:

• What features matter
• Which products fit the budget
• Which options suit the use case
• How the products compare
• Where suitable products are available
• How the customer can complete the purchase

This moves ecommerce from command-based interaction to goal-based help.

Understanding Requirements

The first job of an AI shopping agent is to understand what the person really wants.

This can include:

• Product type
• Intended use
• Budget
• Preferences
• Restrictions
• Time requirements
• Location
• Brand preference

The customer can give these details naturally instead of converting everything into filters.

For example:

“I need a printer for a home office. I print about 100 pages a month, mostly black and white, and I don’t want expensive replacement ink.”

The AI can use this real-world need as the base for finding products.

Research

AI can then help with product research.

Traditional product research often means visiting many websites, reading reviews and comparing specs by hand.

An AI shopping agent can help gather the relevant information and present it in a more usable form.

Instead of reading ten product pages, the shopper might ask:

“What are the important differences between these three printers?”

This cuts down the manual research needed before making a decision.

Discovery

Discovery is the next stage.

The AI can help find products the customer may not know about yet.

This is especially important for retailers with big catalogues.

Customers can’t search for a product they don’t know exists.

An AI system can potentially match a need with products that fit it.

This moves product visibility beyond exact keyword matching.

The important question becomes:

Is there enough reliable information for AI to understand when this product is relevant?

Comparison

Once suitable options are found, AI can help compare them.

A customer might ask:

“Which one is better for a small apartment?”

or:

“Which one offers the best value if I only use it twice a week?”

or:

“Which product has the strongest mix of price, reviews and battery life?”

These are hard questions to solve with basic filters.

They need the system to judge products against the customer’s own situation.

Recommendations

Agentic commerce also changes how recommendations can work.

Traditional ecommerce recommendations often follow patterns like:

Customers also bought

or

Similar products

These systems are useful, but they don’t necessarily understand the customer’s current goal.

An AI recommendation can be built around the current conversation.

For example:

“Based on what you’ve told me, Product B looks closer to your needs because it’s lighter and has longer battery life, while Product A is cheaper but heavier.”

This makes the recommendation feel specific instead of generic.

Purchase Assistance

Agentic commerce becomes most meaningful when AI can help move the shopper from recommendation to action.

In September 2025, OpenAI introduced Instant Checkout and the Agentic Commerce Protocol, developed with Stripe and merchant partners.

OpenAI described ACP as an open standard that lets AI agents, people and businesses work together to complete purchases.

When it’s supported, ChatGPT can pass along the information needed for an order to the merchant, while the merchant stays responsible for things like payment processing, delivery and customer service.

This is an important point.

The AI doesn’t need to replace the retailer.

It can work as a layer between the customer’s intent and the retailer’s existing commerce systems.

Personalisation

Personalisation may end up being one of the strongest parts of agentic commerce.

A traditional product search might treat two customers searching for “running shoes” almost the same.

A conversational system can understand that:

Customer A

• Runs daily
• Trains for marathons
• Prioritises cushioning

while:

Customer B

• Runs twice a week
• Has a $70 budget
• Mostly runs on a treadmill

The same broad search can lead to very different product needs.

Agentic commerce gives retailers a chance to make discovery more relevant to each person’s situation.

The Role of the Agentic Commerce Protocol

OpenAI launched ACP in 2025 as infrastructure that connects AI shopping experiences with merchant systems.

The protocol helps merchants share structured information with ChatGPT and supports commerce workflows without forcing them to rebuild their existing systems.

OpenAI’s current developer documentation describes ACP as the connecting layer between merchants and shoppers in ChatGPT. It can take in structured catalogue information, understand merchant inventory and show products in context.

Structured product feeds can include details such as:

• Product identifiers
• Descriptions
• Images
• Pricing
• Inventory
• Fulfilment information

OpenAI also recommends refreshing feeds regularly, so ChatGPT works with more accurate product information.

This makes agentic commerce partly an AI challenge and partly a data-quality challenge.

AI can only make useful recommendations if it has dependable information to work with.

Why Agentic Commerce Matters

Agentic commerce matters because it could shorten the distance between intent and transaction.

The traditional digital shopping journey often makes shoppers turn their own needs into a series of searches and clicks.

The new model lets AI do more of that translation.

The progression looks like this:

Customer Goal ↓ AI Understands Requirements ↓ Research and Discovery ↓ Product Comparison ↓ Recommendation ↓ Commerce Action

The customer stays in control.

OpenAI’s Instant Checkout framework requires users to confirm purchases instead of letting AI make buying decisions without limits.

That difference matters.

The aim of agentic commerce is not necessarily to remove the customer from the process.

It is to remove unnecessary work from the process.

Search, Recommendation and Checkout Are Starting to Merge

For most of ecommerce history, finding a product and buying it happened in separate places.

A customer might start on Google, move to a retailer, compare products and finally reach checkout.

The usual journey looked like this:

Google → Walmart → Search → Product Pages → Compare → Cart → Checkout

Each stage had a clear job.

Google helped the customer find information.

Walmart helped the customer browse products.

Product pages explained individual items.

The cart collected the selections.

Checkout completed the sale.

AI is starting to blur these lines.

A newer journey may look more like this:

Ask AI → Discover → Compare → Select → Commerce Action

Several stages that once needed different screens can now happen inside one conversation.

Discovery Is Moving Closer to Decision-Making

Search engines have always been good at helping people find pages.

AI assistants are increasingly trying to help people find answers.

Shopping adds one more step: helping people make decisions.

A customer doesn’t always want 100 links for coffee machines.

They may simply want to know:

“Which coffee machine makes sense for my kitchen, budget and daily routine?”

The value shifts from returning more results to cutting out irrelevant choices.

This changes the role of product discovery.

The goal is no longer just:

“Show products.”

It becomes:

“Help me understand which products fit my situation.”

Recommendation Is Becoming Part of Search

In the past, recommendation systems usually appeared after a customer had already landed on a retailer’s site.

For example:

• Recommended for you
• Related products
• Customers also viewed
• Frequently bought together

Conversational AI can move recommendation much earlier.

The customer can get recommendations before they even reach the retailer’s normal search screen.

This creates a new discovery layer between the shopper and the online store.

That layer matters for brands because AI may influence which products make it into the shopper’s list of options in the first place.

A product that never shows up in the AI-assisted research may have fewer chances to be considered later.

Comparison Is Moving Into the Discovery Layer

Comparison is also becoming part of the conversation.

A customer can ask:

“Compare these three based on price, reliability and battery life.”

Then say:

“Remove the most expensive one.”

Then:

“Which of the remaining two is better for travel?”

Each question moves the customer closer to a buying decision.

In a traditional funnel, comparison often happens across different websites and browser tabs.

In an AI-powered journey, discovery and comparison can become part of one interaction.

Checkout Is Moving Closer to Recommendation

The biggest structural change happens when checkout becomes linked with the conversation.

OpenAI’s Instant Checkout launch in September 2025 let eligible purchases be completed without leaving ChatGPT, while merchants kept managing orders, payments and delivery on their own systems.

That reduces the gap between:

Recommendation

and

Transaction

Before:

AI recommendation → Leave AI → Find retailer → Find product → Add to cart → Checkout

The emerging experience:

AI recommendation → Select product → Commerce action

Fewer steps could make conversational commerce especially powerful.

The Funnel Is Becoming More Compressed

Traditional ecommerce marketing often treats the customer journey as a funnel.

Awareness The customer learns about the product.

Consideration The customer researches alternatives.

Comparison The customer evaluates products.

Purchase The customer completes a transaction.

AI can squeeze some of these stages together.

A single conversation might include:

“What do I need?” ↓ “What products fit?” ↓ “Which one is better?” ↓ “How much does it cost?” ↓ “Can I get it tomorrow?” ↓ “I’ll take that one.”

Instead of many sessions across different platforms, much of the decision can happen within the same context.

Discovery Platforms and Shopping Platforms Are Starting to Overlap

This creates an important change in digital commerce.

Before:

• Google was mainly a discovery platform.
• Walmart was mainly a shopping platform.

Conversational commerce makes these roles overlap.

ChatGPT can help with:

• Discovery
• Research
• Understanding products
• Comparison
• Recommendation
• Purchase help

Retailers like Walmart then provide:

• Catalogue depth
• Inventory
• Account relationships
• Loyalty
• Payment
• Delivery
• Customer support

The line between where customers find products and where they buy them is becoming less rigid.

Why This Matters for Brands

The effects reach well beyond Walmart.

A traditional digital marketing plan might focus heavily on:

• Google rankings
• Google Shopping
• Marketplace visibility
• Paid search
• Retail media
• Social ads

These channels are still important.

But AI adds another discovery surface.

Brands may now need to ask:

• Does AI know our product exists?
• Does it understand what the product is good for?
• Can it tell our product apart from competitors?
• Does it have current price and availability information?
• Is there enough trustworthy information for it to recommend the product in the right situation?

This is where AI Visibility turns into a strategic ecommerce issue, not just a technical SEO topic.

The new funnel doesn’t remove traditional search.

It adds one more layer:

Search visibility + Marketplace visibility + AI visibility

Brands that understand all three may be better prepared for a discovery environment that is splitting into many parts.

The Numbers Behind Walmart’s Ecommerce Momentum

Walmart’s AI strategy looks even more significant when you put it next to the size and growth of its ecommerce business.

The company is not adding conversational commerce to a weak digital operation.

It is adding AI to an ecommerce system that is already growing fast across online sales, marketplace activity, delivery and advertising.

The FY2027 numbers show how strong that momentum is.

Q1 FY2027: Ecommerce Growth Picked Up

For the first quarter of FY2027, Walmart reported $177.8 billion in total revenue, up 7.3% year over year.

More important for this case study, global ecommerce sales grew 26%.

Walmart said this growth was led by store-fulfilled pickup and delivery, along with marketplace activity.

Walmart Q1 FY2027 Digital Highlights

Metric	Q1 FY2027 Result
Global ecommerce growth	+26%
Ecommerce share of total net sales	23%
Walmart U.S. ecommerce growth	+26%
Store-fulfilled delivery growth	~45%
Marketplace sales growth	Nearly +50%
Walmart Connect growth*	+44%
Express delivery share	~36% of store-fulfilled orders

*Walmart Connect growth excludes VIZIO where Walmart has noted it.

These figures show that several parts of Walmart’s digital ecosystem were growing together.

Ecommerce wasn’t being driven only by people ordering regular store products online.

Marketplace, delivery and advertising were also growing quickly.

Ecommerce Reached 23% of Walmart Net Sales

One of the most important numbers is how much ecommerce contributes to the whole business.

In Q1 FY2027, ecommerce made up about 23% of Walmart’s total net sales.

That means nearly one dollar out of every four of Walmart’s net sales was already tied to ecommerce.

For a company of Walmart’s size, this makes changes to the digital customer journey strategically important.

Conversational product discovery doesn’t need to replace every traditional ecommerce interaction to matter.

Even influencing a slice of such a large digital business could make AI shopping commercially meaningful.

Walmart U.S. Ecommerce Grew 26%

Walmart U.S. ecommerce sales also rose 26% in Q1 FY2027.

Walmart credited the result to strength in areas including:

• Store-fulfilled delivery
• Advertising
• Marketplace

This shows something important about Walmart’s ecommerce model.

Its online business is not separate from its physical stores.

The stores increasingly work as delivery infrastructure for digital orders.

That gives Walmart an advantage that pure online retailers don’t have in quite the same form.

Store-Fulfilled Delivery Grew Around 45%

Store-fulfilled delivery grew about 45% in Q1 FY2027.

That number gets more interesting when you think about conversational commerce.

AI can help a customer pick a product, but speed and availability still decide whether the purchase actually happens.

If Walmart can combine:

AI recommendation

with

local inventory

and

fast delivery

then conversational shopping becomes more useful.

The customer isn’t just getting a product suggestion.

They may increasingly get an answer to a much more practical question:

“What should I buy, and how quickly can I get it?”

Marketplace Sales Grew Nearly 50%

Walmart’s U.S. marketplace sales rose nearly 50% in Q1 FY2027, which Walmart described as its strongest marketplace performance in ten quarters.

Marketplace growth adds to the product selection.

More products can help shoppers, but they can also make discovery harder.

A bigger catalogue creates a stronger need for tools that help customers find the most relevant options.

This is where conversational AI may be especially useful.

As product choice grows, the challenge shifts from:

“Do we have enough products?”

to:

“Can customers easily find the right product?”

AI-assisted discovery could help solve that.

Walmart Connect Grew 44%

Walmart’s advertising business was also growing fast.

In Q1 FY2027, total Walmart U.S. advertising grew 36%, including a 44% jump in Walmart Connect, excluding VIZIO.

This is important because ecommerce, advertising and product discovery increasingly affect one another.

Walmart already runs a retail media business that helps brands reach shoppers inside its ecosystem.

AI adds another possible discovery layer.

OpenAI states that its current ChatGPT shopping product results are ranked on relevance and not influenced by merchant payments. Still, the wider retail world is clearly heading toward a setup where AI-assisted discovery exists next to traditional retail advertising.

For brands, that makes the difference between paid visibility and AI recommendation visibility more and more important.

Q2 FY2027: Digital Momentum Continued

Walmart’s Q2 FY2027 results showed the digital growth trend carrying on.

Walmart U.S. generated $125.2 billion in net sales for the quarter, up 3.5% year over year.

Ecommerce remained one of the strongest areas of growth.

Walmart Q2 FY2027 Digital Highlights

Metric	Q2 FY2027 Result
Walmart U.S. ecommerce growth	+24%
Store-fulfilled delivery growth	~43%
Marketplace sales growth	+52%
Total U.S. advertising growth	+38%
Walmart Connect growth*	+43%
Express delivery share	~37% of store-fulfilled orders

*Excluding VIZIO, as reported by Walmart.

The key point is not just that ecommerce kept growing.

Several nearby businesses grew with it.

Marketplace Growth Sped Up to 52%

U.S. marketplace sales rose 52% in Q2 FY2027, compared with nearly 50% growth in the previous quarter.

That growth adds more sellers, products and variations to Walmart’s digital ecosystem.

For shoppers, that means more choice.

For AI, it means more information to understand and organise.

For merchants, it raises the importance of giving clear and accurate product data.

In an AI-assisted shopping environment, weak product information may become a visibility disadvantage.

If the system can’t tell exactly what a product is, who it’s made for or how it differs from alternatives, recommending it becomes harder.

Store-Fulfilled Delivery Stayed Strong at Around 43%

Store-fulfilled delivery grew about 43% in Q2 FY2027.

The small slowdown from roughly 45% in Q1 is still very fast growth.

Even more notable, Walmart said about 37% of store-fulfilled orders were express deliveries completed in under three hours.

That is up from around 36% in Q1.

This gives Walmart a strong link between digital discovery and physical delivery.

A product recommendation becomes more attractive when the system can eventually combine:

Relevant product + Current inventory + Competitive price + Fast delivery

This is the kind of setting where AI commerce can become truly useful and not just a novelty.

Walmart Connect Kept Growing at 43%

Walmart Connect grew 43% in Q2 FY2027, excluding VIZIO.

This continued strength shows that brands increasingly see Walmart as more than a retailer. They see it as a marketing platform too.

Walmart’s digital ecosystem now brings together several valuable parts:

• Ecommerce
• Marketplace
• Retail Media
• Stores
• Delivery
• Membership
• AI Discovery

Each part makes the others stronger.

Walmart’s Ecommerce Momentum in One View

The move from Q1 to Q2 FY2027 can be summed up like this:

Digital Metric	Q1 FY2027	Q2 FY2027
Walmart U.S. ecommerce	+26%	+24%
Store-fulfilled delivery	~45%	~43%
Marketplace	Nearly +50%	+52%
Walmart Connect*	+44%	+43%
Express delivery share	~36%	~37%

*Excluding VIZIO.

This table shows why Walmart is such a useful case study for AI commerce.

The company already has:

• Fast ecommerce growth
• A widening marketplace selection
• Growing retail advertising
• Physical stores acting as delivery hubs
• Faster delivery
• Large-scale customer relationships

ChatGPT and agentic commerce are being added on top of this existing infrastructure.

Why These Numbers Matter for the AI Case Study

AI alone doesn’t build a strong ecommerce business.

A successful AI shopping experience still depends on what happens after the recommendation.

The retailer needs:

• Accurate product information
• Competitive pricing
• Inventory
• Payments
• Customer accounts
• Delivery
• Returns
• Support

Walmart already runs these systems at a massive scale.

That is why the company’s AI strategy deserves attention.

ChatGPT can improve the front end of product discovery, while Walmart’s infrastructure handles the commercial reality behind the recommendation.

This creates a powerful mix:

AI understands demand.

Walmart provides supply.

The commerce layer connects the two.

The wider lesson for ecommerce businesses is that success in AI shopping is unlikely to come from AI alone.

It will depend on how well a company combines:

AI Visibility + Product Data + Customer Experience + Inventory + Fulfilment + Trust

Walmart’s numbers show that the company is building conversational commerce on top of an already growing digital ecosystem.

That makes its partnership with OpenAI more than an experiment with a new interface.

It is an attempt to link one of the world’s largest retail operations with a new way for consumers to discover, evaluate and eventually buy products.

AI Is Already Producing Operational Gains Inside Walmart

Walmart’s AI story is much bigger than its partnership with ChatGPT.

Conversational shopping may be the most visible part of Walmart’s AI strategy, but the company was already using artificial intelligence across many areas of its business before it launched AI-powered shopping experiences.

This matters because it shows ChatGPT is not a one-off experiment.

It is part of a much wider change in how Walmart manages products, serves customers, supports employees and runs at scale.

According to Walmart, AI has already helped the company improve areas such as:

• Product catalogue management
• Fashion production
• Customer support
• Internal productivity
• Shopping experiences

Walmart said its use of AI has helped cut some fashion production timelines by up to 18 weeks and reduce customer-care resolution times by up to 40%.

These are major operational gains.

They show how AI can create value long before a customer starts talking to a shopping assistant.

AI Is Helping Walmart Improve Its Product Catalogue

Product information is one of the most important foundations of ecommerce.

For a large retailer, keeping a catalogue in order is very complex.

Each product may include details such as:

• Product name
• Category
• Description
• Images
• Specifications
• Sizes
• Colours
• Brand
• Price
• Availability
• Seller details

When millions of products are involved, even small improvements in catalogue management can have a large operational effect.

Walmart has said it is using AI to improve its product catalogue.

This is especially relevant to conversational commerce.

AI shopping depends heavily on accurate product information.

If product data is unclear, incomplete or out of date, an AI system may struggle to know when a product is relevant.

Better catalogue quality therefore supports both traditional ecommerce and new AI discovery.

AI Has Cut Some Fashion Production Timelines by Up to 18 Weeks

Walmart also reported that AI has helped shorten certain fashion production timelines by up to 18 weeks.

This shows a different kind of AI value.

Instead of customer-facing recommendations, AI here is helping speed up how fast products move through parts of the production process.

For a retailer, faster production can mean:

• A quicker response to changing demand
• Faster product launches
• More efficient planning
• Fewer operational delays
• Better alignment between trends and available products

This is a good reason to think of AI as infrastructure and not just a customer-facing feature.

The technology can shape what happens before a product ever appears online.

Customer-Care Resolution Times Have Improved by Up to 40%

Customer service is another area where Walmart says AI has produced measurable results.

The company reported cuts of up to 40% in customer-care resolution times.

This matters because ecommerce doesn’t end when a customer clicks “buy.”

Customers may still need help with:

• Delivery
• Returns
• Refunds
• Product questions
• Account problems
• Order changes

Faster resolution can improve the overall shopping experience.

It can also reduce friction after the purchase.

This ties in with agentic commerce.

AI may eventually support several stages of the customer relationship:

• Before purchase: discovery and comparison
• During purchase: selection and checkout help
• After purchase: service and support

That makes AI potentially useful across the whole commerce lifecycle.

AI Is Also Becoming an Internal Business Tool

Walmart’s AI strategy isn’t limited to customer-facing technology.

The company has also been widening AI access for employees, including rolling out ChatGPT Enterprise to teams and running programmes to improve AI skills among associates.

This suggests Walmart sees AI as a general business capability, not a single product.

It can support:

• Operations
• Employees
• Customer service
• Product data
• Shopping
• Decision-making

The bigger picture is clear.

Walmart isn’t adding ChatGPT to a traditional retail company and calling it an AI strategy.

It is gradually adding AI across many layers of the organisation.

Conversational commerce is just one of the most visible customer-facing results of that wider change.

Why Conversational Shopping Could Beat Keyword Search for Complex Purchases

Traditional keyword search is unlikely to disappear.

For many shopping tasks, it is still extremely efficient.

If a customer already knows exactly what they want, typing the exact product name into a search bar can be faster than holding a conversation.

For example:

“Apple AirPods Pro”

or:

“55-inch Samsung QLED TV”

These are simple, specific searches.

Conversational shopping becomes more useful when the purchase is less clear.

The customer may know the problem but not the product.

That is where AI can add more value.

Vague Needs

Think about someone saying:

“I need something to make my home office more comfortable.”

This is hard for a keyword search engine.

What does the customer need?

• A chair?
• A desk?
• A monitor stand?
• A lamp?
• A footrest?
• A keyboard?
• A cushion?

Traditional search asks the shopper to decide which product category to explore first.

Conversational AI can ask questions.

For example:

“Are you mainly uncomfortable because of your chair, desk height, lighting or computer setup?”

This helps turn a vague problem into a clearer shopping requirement.

Multiple Requirements

Many purchases involve several conditions at once.

For example:

“I need a laptop under $900 that is light enough for travel, has good battery life and can handle basic video editing.”

This includes:

• Budget
• Weight
• Battery
• Performance
• Use case

Traditional ecommerce can handle some of these through filters.

But the customer still has to know which specifications matter.

Conversational AI can potentially translate the real-world requirement into product criteria.

That makes it especially useful when several priorities need to be balanced.

Gift Shopping

Gift buying is another strong use case.

A shopper may know the person but not know what to buy.

For example:

“I need a birthday gift for my father. He likes cooking and outdoor grilling, and I want to spend around $75.”

A traditional search might start with:

“Gifts for dads”

That can produce a very broad list.

A conversation can keep narrowing the options:

“He already has a grill.”

“He likes practical gifts rather than decorative ones.”

“Nothing that takes up too much space.”

Each reply improves the recommendation.

Shopping starts to feel like guided discovery.

Meal Planning

Meal planning is another case where plain product search gets clumsy.

Imagine asking:

“I need five easy dinners for two people this week. Keep the total grocery budget below $80 and avoid dairy.”

A traditional shopping process may need you to:

1. Find recipes.
2. Write an ingredient list.
3. Search for each item.
4. Check quantities.
5. Compare prices.
6. Add everything to the basket.

An AI shopping assistant could potentially link these steps together.

The customer isn’t really searching for one product.

They are trying to finish a bigger task.

This is exactly the kind of problem where agentic commerce becomes more interesting.

Product Comparison

Comparison is another area where conversational shopping can have an edge.

Suppose a shopper is choosing between three vacuum cleaners.

A product page may give specifications.

A comparison page may give a table.

But the shopper might really want to know:

“Which one is best for a small home with two pets and mostly hard flooring?”

That question mixes several product attributes with a personal situation.

ChatGPT’s shopping research is specifically designed for decisions with comparisons, trade-offs and several constraints. Users can share preferences and budget details and refine recommendations through follow-up questions.

This is where conversational discovery can beat simply showing more product listings.

Budget Constraints

Budget is rarely as simple as a maximum price.

A shopper may say:

“I want to stay under $300, but I can spend another $30 if it gives me much better battery life.”

A normal price filter struggles to capture that preference.

The customer would have to choose either:

Maximum: $300

or:

Maximum: $330

But the real instruction is more nuanced.

The shopper is willing to pay more only if the added value is meaningful.

Conversational AI is better at understanding these kinds of trade-offs.

Household Restocking

Routine purchases are another possible use case.

Households regularly buy again:

• Cleaning products
• Toiletries
• Food
• Pet supplies
• Paper products
• Household essentials

Over time, AI shopping assistants may help reduce the effort in these repeat decisions.

Instead of finding every item by hand, the customer could ask:

“Help me restock the household essentials I usually need for the next two weeks.”

The value comes from turning many small shopping actions into one bigger task.

Personalised Recommendations

Conversational shopping can also use context to make recommendations more specific.

OpenAI says ChatGPT shopping results can take into account a user’s query and the context of the conversation. If relevant context is available, it can affect which products are shown.

For example, two people might ask:

“Recommend a laptop.”

But one may care about gaming performance, while the other mainly cares about battery life and portability.

A conversational system can give different recommendations based on those priorities.

Keyword Search Still Has an Important Role

It would be wrong to say AI is completely replacing ecommerce search.

Traditional search is still useful for:

• Exact products
• Known brands
• Repeat purchases
• Specific model numbers
• Quick price checks
• Simple purchases

The more realistic change is that AI is becoming an extra product discovery layer.

Customers may use traditional search when they know exactly what they want.

They may use conversational AI when they need help deciding.

So the future of ecommerce is unlikely to be:

Search OR AI

It is more likely to be:

Search + AI + Marketplace Discovery + Recommendations

Different interfaces will serve different stages of the customer journey.

Walmart’s Bigger Advantage: Data + Inventory + Fulfilment + AI

AI recommendations are only useful when they connect to real commerce.

An AI assistant may spot the perfect product, but the recommendation loses its value if:

• The product is out of stock
• The price is wrong
• Delivery is too slow
• The retailer can’t fulfil the order
• The checkout experience fails

This is where Walmart may have an important structural advantage.

Its strength is not AI alone.

It is the mix of several assets:

• AI
• Product Catalogue
• Marketplace Sellers
• Customer Relationships
• Physical Stores
• Inventory
• Fast Fulfilment

Together, these pieces can turn conversational recommendations into real sales.

AI Provides the Intelligence Layer

AI can help understand what the customer wants.

For example:

“I need inexpensive patio furniture for a small apartment balcony.”

The AI can pick up on:

• Product category
• Size limits
• Price sensitivity
• Likely use
• Customer priorities

But understanding the request is only the first step.

The system still needs relevant products.

Walmart’s Product Catalogue Provides Choice

Walmart already runs a large product catalogue across many categories.

That gives AI a wide range of options to work with.

A large catalogue becomes even more valuable when AI can help customers move through it.

In traditional ecommerce, adding more products can sometimes add more complexity.

Customers may struggle to find the right item among thousands of similar ones.

Conversational discovery could turn catalogue size from a navigation problem into a recommendation advantage.

Instead of asking customers to search the whole catalogue themselves, AI can help narrow the selection.

Marketplace Sellers Expand the Assortment

Walmart’s marketplace adds even more product variety.

Third-party sellers can offer extra:

• Brands
• Price points
• Product variations
• Niche items
• Categories

This increases the number of possible solutions available to the shopper.

But a bigger marketplace also makes product data more important.

If AI has to compare similar products from many sellers, the information needs to be accurate and easy to understand.

That makes product feeds, structured attributes and catalogue quality more and more important.

Customer Relationships Add Context

Walmart also has direct relationships with a very large number of customers.

Accounts, membership programmes and shopping history can help create a more connected retail experience where users choose to link or use those features.

Walmart’s ChatGPT integration is especially interesting because OpenAI has said the Walmart experience supports functions including account linking and loyalty.

That opens the door to connecting conversational discovery with an existing customer relationship.

Instead of AI working completely apart from the retailer, the experience can become part of a wider shopping ecosystem.

Physical Stores Become Digital Infrastructure

Walmart’s physical stores were once seen mainly as places where people shopped in person.

In an omnichannel model, they can also work as fulfilment hubs.

This is important for AI commerce.

A customer might ask:

“I need a phone charger today.”

The best recommendation isn’t necessarily the product with the highest review score.

It may be the suitable product that is:

• Available nearby
• Within budget
• Compatible with the customer’s device
• Available for fast delivery

So physical inventory becomes part of the value of a recommendation.

An AI system gets more useful when it can eventually link product relevance with real-world availability.

Inventory Makes Recommendations Actionable

Inventory data answers a key question:

Can the customer actually buy the product?

Without inventory information, product recommendations can become frustrating.

For example:

1. AI recommends a product.
2. Customer clicks.
3. Product is unavailable.
4. The customer starts again.

That breaks the experience.

Better links between AI and commerce infrastructure can reduce this kind of friction.

OpenAI has said its Agentic Commerce Protocol is growing to support richer product information, with longer-term features including local availability and estimated arrival times.

For retailers like Walmart, that direction is strategically important.

Fast Fulfilment Turns Advice Into Convenience

The final advantage is delivery.

Walmart’s store network already supports large-scale delivery and pickup.

When fast delivery is combined with AI discovery, the offer gets stronger.

The customer is no longer just asking:

“What should I buy?”

They can increasingly care about:

“What should I buy that I can receive today?”

This changes the value of the recommendation.

A useful commerce system has to understand both:

Product relevance

and

Commercial reality

Walmart has infrastructure for both.

Walmart’s AI Advantage Is the Combination, Not One Technology

So the competitive advantage shouldn’t be understood as:

Walmart has AI.

Many companies have access to AI technology.

The stronger advantage is:

Walmart can connect AI with a large retail ecosystem.

The combination looks like this:

Customer Intent ↓ AI Understanding ↓ Large Product Catalogue ↓ Marketplace Selection ↓ Inventory ↓ Account and Loyalty ↓ Payment ↓ Fast Fulfilment

This is what can make conversational commerce commercially useful instead of merely interesting.

The AI recommendation gets tied to a real ability to deliver the product.

AI Discovery Changes the Meaning of “Being Visible Online”

For most of the digital era, online visibility was mainly defined by search engines, advertising platforms and marketplaces.

A business might ask:

“Do we rank on Google?”

“Are our Shopping ads showing up?”

“Where do we rank on Amazon or Walmart?”

“Are our paid campaigns reaching the right customers?”

These questions still matter.

But AI-driven discovery adds another layer.

Businesses now need to think about:

“When a customer asks AI what to buy, does our brand appear?”

This changes what digital visibility means.

Traditional Online Visibility

Traditional ecommerce visibility has usually depended on channels such as:

Google Rankings Brands compete for organic search positions on commercial and informational keywords.

For example:

“best running shoes for beginners”

Paid Search Businesses bid for ad visibility when users search relevant terms.

Shopping Ads Product feeds are used to show products with images, prices and merchant details.

Marketplace Rankings Retailers and sellers compete for visibility inside platforms such as Walmart and other large marketplaces.

These channels are still major sources of ecommerce traffic and revenue.

AI doesn’t make them irrelevant.

It adds another discovery environment.

Emerging AI Visibility

Consumers can now begin their research inside AI systems.

Instead of typing:

“best office chair under $300”

into a normal search engine, someone might ask:

“I work from home eight hours a day and need a comfortable office chair under $300. What should I consider?”

The response may include:

• Product suggestions
• Comparisons
• Buying advice
• Brand mentions
• Merchant options

This brings in several new types of visibility.

ChatGPT Recommendations Brands and products can show up when ChatGPT decides they fit a shopping request.

Gemini Answers Consumers may use Google’s Gemini ecosystem to explore and research products in conversation.

Google AI Overviews AI-generated search experiences can change how information is summarised before users click regular organic results.

AI Shopping Results AI interfaces can show product information directly inside a conversation.

Conversational Recommendations A product may show up because it suits the discussion, not because the user typed its exact keyword.

AI-Generated Product Comparisons AI may compare products based on:

• Price
• Features
• Reviews
• Use case
• Customer priorities

That means visibility depends less on one fixed ranking page.

From Keyword Visibility to Contextual Relevance

Traditional SEO has focused heavily on keywords.

For example:

• “running shoes”
• “best running shoes”
• “women’s running shoes”

These are still useful search signals.

But conversational queries can be much more specific.

For example:

“I run three times a week on pavement and sometimes get sore feet. I need comfortable shoes under $120.”

A product could be relevant to that request even if the customer never typed its brand name.

This changes the visibility challenge.

Brands need to make it easy for AI systems to understand:

• What the product is
• Who it’s for
• What problem it solves
• How much it costs
• Its important features
• Where it’s available
• How it differs from alternatives

That is a broader job than simply targeting a keyword.

AI Visibility Is About Being Understandable

Being indexed is not the same as being understood.

A brand can have hundreds of web pages and still give unclear information about:

• Its products
• Use cases
• Customers
• Differences from competitors
• Pricing
• Availability

AI discovery makes clear entity and product information more important.

The goal becomes:

Make it easy for AI systems to correctly understand what your brand and products stand for.

This can involve:

• Strong product descriptions
• Clear website structure
• Structured data
• Accurate merchant feeds
• Consistent brand information
• Helpful comparison content
• FAQs
• Reviews
• Authoritative third-party mentions

From Ranking to Recommendation

Traditional search optimisation often asks:

“Can we rank?”

AI visibility adds a second question:

“Can we be recommended?”

These are not the same goal.

A page can rank because it’s relevant to a certain query.

A recommendation may need a deeper understanding of:

• Product suitability
• User context
• Price
• Features
• Reviews
• Merchant quality
• Availability

So the move from ranking to recommendation raises the standard for product information.

AI Visibility Does Not Replace SEO

It’s important not to present this as the end of SEO.

SEO is still important because:

• Search engines still drive discovery.
• AI systems can use information found across the web.
• Strong websites help build product and brand understanding.
• Structured content helps both search engines and AI systems.

A better model is:

SEO + AEO + GEO + Marketplace Optimisation + AI Visibility

These areas overlap more and more.

The goal is not to give up on traditional visibility.

It is to widen the definition of visibility.

For ecommerce brands, being visible online may increasingly mean showing up across:

Search Engines + Marketplaces + Ads + AI Answers + AI Recommendations

How Does ChatGPT Decide Which Products to Show?

For brands trying to understand AI shopping, one of the biggest questions is:

“How does ChatGPT decide which products appear?”

There is no public secret formula that merchants can play with to guarantee a recommendation.

However, OpenAI does explain several factors that go into product selection.

According to OpenAI, a product may appear when ChatGPT considers it relevant to the user’s intent and context.

OpenAI also clearly states that shopping product results are not ads and are not chosen because of OpenAI partnerships. Advertising appears separately from organic shopping results.

This difference is important.

A merchant can’t just assume that paying for advertising guarantees a spot in organic product recommendations.

User Intent Comes First

The customer’s request plays the central role.

Take two queries:

“Affordable wireless headphones.”

and:

“Noise-cancelling headphones for long flights under $150.”

The second request has much more context.

ChatGPT can take into account:

• Product type
• Intended use
• Required feature
• Budget

OpenAI says product selection can change depending on what the user needs. If a shopper mentions a budget, for example, price can become more important in deciding which results are relevant.

This means merchants should think beyond single keywords.

Products need enough information for AI to understand when they fit a real-world need.

Context Can Influence Product Relevance

ChatGPT can also use the context around the request.

OpenAI says shopping results may consider the user’s query together with available context, such as relevant preferences.

For example, if a conversation shows that the user:

• Has a limited budget
• Prefers lightweight products
• Needs something for travel

then those preferences can affect which products make sense.

This is fundamentally different from exact-match keyword search.

A product can be relevant because of what it does, not just because its title contains certain words.

Accurate Product Descriptions Matter

If an AI system is going to understand a product, the description has to clearly say what the product is.

Weak description:

“Premium quality product with amazing performance.”

This tells the system almost nothing.

Better information might explain:

• What the product does
• Who it’s for
• Major features
• Materials
• Compatibility
• Important dimensions
• Intended use

Clear product descriptions improve semantic understanding.

The goal should not be keyword stuffing.

It should be product clarity.

Price Needs to Be Accurate

Price is often part of a shopper’s request.

OpenAI says ChatGPT may consider price when judging product relevance, especially when the user states a budget.

For example:

“Find me a coffee machine under $150.”

A product priced at $300 should normally be less relevant to that request.

This makes accurate, current pricing information important.

OpenAI also notes that prices can sometimes take a while to update when a merchant’s pricing changes, which is one reason direct and regularly refreshed merchant data can matter.

Availability Matters

A product recommendation is less useful when the product can’t be bought.

OpenAI says merchant ranking can consider factors including availability, along with price, quality and whether the merchant is the maker or primary seller.

For ecommerce brands, inventory information therefore becomes part of AI visibility.

It isn’t enough for an AI system to understand the product.

It also needs reliable commerce information.

Product Attributes Help AI Match Specific Needs

Product attributes help tell similar products apart.

For example, a laptop may include:

• Screen size
• RAM
• Storage
• Weight
• Processor
• Battery specifications

Shoes may include:

• Size
• Material
• Intended activity
• Cushioning
• Width
• Colour

These attributes become valuable when users make detailed requests.

For example:

“I need a lightweight 14-inch laptop with at least 16GB RAM.”

Clear attributes make it easier to match products to requests like this.

Images Support Visual Product Discovery

ChatGPT shopping is getting more visual.

OpenAI’s March 2026 shopping update added richer visual browsing and product comparison experiences.

That makes product images another important commerce asset.

Good images help users judge products visually and support richer product cards.

For merchants, this reinforces the importance of:

• Clear product photography
• Accurate images
• Consistent variants
• Useful visual details

The product experience is no longer text-only.

Reviews Can Influence Product Understanding

OpenAI says ChatGPT can use reviews and ratings as part of the information available when helping users compare products. It can also create summaries that highlight common positives and negatives found in public reviews.

This means customer feedback adds another layer of product information.

For example, the product specifications may say a pair of shoes is lightweight.

Reviews may show whether customers consistently find them comfortable on long walks.

Those are different kinds of information.

Both can help the customer decide.

OpenAI also notes that review summaries and ratings shown in ChatGPT are not independently verified by OpenAI, so users should still judge them carefully.

Merchant Data Can Improve Accuracy

OpenAI says product information may come from structured metadata supplied by first-party and third-party providers, as well as other third-party content.

Merchants can also give product feeds to OpenAI.

Direct feeds can help provide more complete and current information about products.

This matters for:

• Price
• Product descriptions
• Availability
• Variants
• Merchant information

The better the data, the easier it is for AI systems to represent the product accurately.

Structured Product Information Is Becoming More Important

Structured information gives systems a clearer way to understand commerce data.

Instead of forcing an AI system to work everything out from marketing copy, structured feeds can share information in defined fields.

For example:

• Product: Running shoe
• Brand: Example Brand
• Price: $89
• Size: 7–12
• Colour: Black
• Use: Road running
• Availability: In stock

This makes product information easier to read in a consistent way.

For ecommerce brands, structured commerce data may become as strategically important for AI discovery as product feeds already are for other shopping platforms.

Brand and Entity Clarity Matter

AI systems also need to correctly understand how these connect:

• Brand
• Product
• Manufacturer
• Merchant
• Product category

Confusing brand information can make this harder.

For example, a business should ideally use the same brand name across:

• Website
• Product pages
• Merchant feeds
• Structured data
• Marketplace listings
• Social profiles
• Authoritative third-party sources

This helps build a clearer entity footprint.

It doesn’t guarantee a ChatGPT recommendation.

But it reduces confusion about who the brand is and what it sells.

What Merchants Should Focus On

Instead of hunting for a hidden ChatGPT ranking trick, merchants should focus on the basics that make products easier to understand and evaluate:

Accurate Product Description Clearly explain what the product is and what it is designed to do.

Current Price Make sure pricing data is accurate and updated often.

Availability Keep inventory information current.

Detailed Attributes Provide useful specifications and variant details.

Strong Product Images Use clear, accurate visuals.

Customer Reviews Build genuine customer feedback.

Structured Product Data Use well-organised product and merchant information.

Consistent Brand Information Make brand identity clear across digital sources.

The strategic lesson is simple:

AI visibility is not about tricking an algorithm. It is about reducing uncertainty.

The easier it is for an AI system to understand:

• what your product is,
• who it’s for,
• what it costs,
• where it’s available,
• and why it may fit a particular need,

the stronger the base for AI-driven product discovery becomes.

That is one of the most important lessons ecommerce brands can take from the shift toward ChatGPT shopping and agentic commerce.

SEO vs AEO vs GEO vs AI Commerce Optimisation

For many years, ecommerce visibility was discussed mainly through SEO.

The basic question was:

“Can customers find our website when they search on Google?”

That question still matters.

But product discovery is getting more complex.

Customers can now search through traditional search engines, ask questions inside AI assistants, use AI-generated search experiences, compare products in conversation and, in some cases, go straight from an AI recommendation to a purchase.

This means brands may need to think about visibility across several connected areas:

SEO + AEO + GEO + AI Commerce Optimisation

These terms overlap, but each one solves a different problem.

What Is SEO?

Search Engine Optimisation, or SEO, focuses on improving visibility in traditional search engines.

A retailer might want to show up when someone searches:

“best office chair for home working”

or:

“women’s running shoes under $100”

SEO usually covers areas such as:

• Keyword research
• Website architecture
• Content quality
• Internal linking
• Page speed
• Technical SEO
• Product pages
• Structured data
• Backlinks
• Brand authority

The goal is usually to help search engines understand a page and improve its chance of appearing for relevant searches.

For ecommerce brands, SEO is still extremely important because traditional search remains a major way people find products.

However, the search experience itself is changing.

Google’s AI-powered search products are becoming more conversational. Google said in 2026 that AI Overviews had reached more than 2.5 billion monthly active users, while AI Mode had passed 1 billion monthly users.

This means SEO now sits inside a wider AI discovery environment.

What Is AEO?

Answer Engine Optimisation, or AEO, focuses on making information easier for systems to understand and use when answering questions.

The difference is easy to see.

Traditional SEO may target:

“best running shoes for beginners”

AEO may prepare information for a question like:

“What type of running shoe should a beginner choose for short road runs?”

The user isn’t necessarily looking for ten blue links.

They may want a direct answer.

So AEO puts strong emphasis on:

• Clear answers
• Helpful explanations
• FAQs
• Well-structured content
• Product specifications
• Definitions
• Comparisons
• User-focused information
• Structured data

The goal is to make information easy to interpret when an answer engine has to explain something clearly.

What Is GEO?

Generative Engine Optimisation, or GEO, focuses on improving how brands, products and information are understood and shown in generative AI environments.

These environments may include:

• ChatGPT
• Gemini
• Google AI experiences
• Perplexity
• Other generative discovery tools

GEO goes beyond ranking a single webpage.

A brand may want AI systems to correctly understand:

• Who the company is
• What products it sells
• What industries it serves
• What makes its products different
• Which customers its products are relevant to
• Whether reliable sources back up its claims

This brings in the idea of entity visibility.

For example, an ecommerce company may rank well for several keywords but still have weak brand visibility across AI systems.

An AI assistant may know the product category but not confidently link that brand with it.

GEO aims to close that gap.

What Is AI Commerce Optimisation?

AI Commerce Optimisation goes one step further.

It focuses specifically on making products and merchant information suitable for AI-driven discovery and commerce experiences.

This involves much more than writing good blog content.

AI-powered shopping systems may need reliable information about:

• Product identity
• Product descriptions
• Price
• Availability
• Variants
• Images
• Seller information
• Brand
• Product attributes
• Fulfilment information

OpenAI’s current product-feed documentation says feeds help ChatGPT show products using structured catalogue data with accurate pricing, availability and seller context. OpenAI also says better feed quality can improve discovery relevance and reduce purchasing friction.

This makes AI Commerce Optimisation partly a marketing discipline and partly a data-management discipline.

The Four Areas Work Together

It would be a mistake to treat SEO, AEO, GEO and AI Commerce Optimisation as completely separate strategies.

They increasingly support one another.

A strong ecommerce brand may need:

SEO So customers can find products through traditional search.

AEO So questions about the products can be answered clearly.

GEO So generative AI systems understand the brand and its relevance.

AI Commerce Optimisation So product information can work well in AI-powered shopping environments.

The wider strategy becomes:

• Be searchable.
• Be understandable.
• Be trusted.
• Be discoverable in AI.
• Be ready for commerce.

That is a much wider definition of ecommerce visibility than traditional SEO alone.

Why Product Feeds May Become More Important in the AI Era

Product feeds have been part of ecommerce marketing for years.

Retailers use them for channels such as:

• Shopping platforms
• Marketplaces
• Advertising systems
• Product comparison tools

The AI commerce era could make these feeds even more important.

The reason is simple.

AI systems need dependable product information before they can confidently match a product with a customer’s request.

OpenAI says product feeds help ChatGPT access accurate and current catalogue information for product discovery and checkout experiences.

A strong product feed effectively tells the system:

• What the product is.
• What it costs.
• Whether it’s available.
• Who sells it.
• Which version of the product the customer is looking at.

Product Titles Need to Be Clear

The product title is one of the most basic signals.

A weak title might be:

“Premium Model X200.”

Unless the brand is already well known, this gives very little context.

A clearer title might explain:

Brand + Product Type + Important Attribute + Model

The goal shouldn’t be to add unnecessary keywords.

It should be to remove confusion.

An AI system should be able to tell what the product actually is.

Product Descriptions Add Context

The description explains the product beyond its title.

A useful description should answer questions such as:

• What does this product do?
• Who is it for?
• What are its main features?
• What problem does it solve?
• How is it normally used?
• Are there important limits?

This information is especially valuable when customers search in conversation.

For example:

“I need a lightweight vacuum for a small apartment with mostly hard floors.”

A useful product description can help the AI judge whether a particular vacuum fits that need.

Category Information Helps Build Product Context

Categories tell systems how a product fits within the wider catalogue.

For example:

Home & Garden → Kitchen → Coffee Makers → Espresso Machines

This gives clearer context than just calling the product:

“Coffee Machine.”

Strong categorisation helps AI tell related but different products apart.

Price Must Stay Current

Price is one of the most common shopping limits.

Customers regularly ask for:

“under $100”

“around $500”

or:

“best option for the money.”

OpenAI’s feed documentation specifically requires current pricing information and recommends updating prices when they change.

Wrong prices can create poor customer experiences.

Imagine:

1. AI shows a product at $89.
2. Customer clicks.
3. The actual price is $119.

Trust drops immediately.

Fresh product feeds help reduce this problem.

Availability Is Equally Important

A great recommendation isn’t useful if the product is out of stock.

Availability therefore becomes a key commerce signal.

OpenAI’s current product-feed specification lists availability as required product information for supported feed formats.

This means merchants increasingly need to think about AI visibility and inventory accuracy together.

AI commerce can’t work well with stale catalogue data.

Variant Information Reduces Confusion

Many products come in multiple variants.

For example, a shoe may come in:

• Several sizes
• Different colours
• Different widths

A laptop may include:

• Different RAM options
• Different storage levels
• Different processors

AI needs to understand these differences correctly.

Otherwise, it may recommend a product that looks suitable but isn’t available in the customer’s required setup.

Clear variant information lowers this risk.

Product Identifiers Improve Accuracy

Product identifiers can help systems tell similar products apart.

These may include:

• SKU
• GTIN
• UPC
• MPN
• Internal product IDs

Identifiers reduce confusion, especially when the same product appears across several merchants or marketplaces.

This becomes more important when AI systems need to compare product information from different sources.

Product Images Matter in Visual AI Shopping

AI shopping is becoming more visual.

OpenAI’s March 2026 shopping update introduced richer visual product browsing and side-by-side comparison experiences.

Images therefore stay an important part of product data.

Useful product imagery should:

• Clearly show the item
• Match the actual variant
• Avoid misleading visual differences
• Give useful product context
• Keep good quality

OpenAI’s supported product-feed format currently requires an image link as part of core product information.

Reviews Add a Customer Experience Layer

Product feeds show what merchants say about their products.

Reviews add another view:

What customers experienced after buying them.

Reviews can give useful information about:

• Quality
• Durability
• Comfort
• Ease of use
• Real-world performance
• Common problems

For AI systems helping customers compare products, these signals can add useful context next to structured product specifications.

Merchant Data Matters Too

Customers aren’t only choosing products.

They may also be choosing where to buy them.

OpenAI says ChatGPT can consider merchant-related information such as:

• Availability
• Price
• Quality
• Whether the merchant is the maker or primary seller

when showing merchant options.

This means product visibility and merchant credibility may become more and more connected.

Product Feeds Could Become Part of AI Visibility Strategy

In the past, many businesses saw product feeds mainly as technical infrastructure.

The AI era may turn them into a visibility asset.

A high-quality feed helps answer important questions:

• What is this product?
• Who sells it?
• How much does it cost?
• Is it available?
• What version is available?
• Is this data current?

OpenAI explicitly states that merchants can provide feeds to help ChatGPT represent products more accurately and completely.

This points to an important ecommerce principle:

Content tells AI why a product matters.

Structured feeds tell AI exactly what the product is right now.

Strong AI commerce strategies may need both.

The Rise of the “AI-Recommended Brand”

For years, one of the most common questions in digital marketing has been:

“Do we rank on Google?”

That question built entire industries around search optimisation.

But another question is now becoming important:

“Will AI recommend our product when somebody asks what to buy?”

This is a different challenge.

Ranking and recommendation are not the same thing.

A website may rank for a keyword because the page fits that search.

An AI recommendation may need to weigh much broader context.

For example:

“I need comfortable walking shoes for a holiday in Europe. I will walk several miles each day and want to stay below $120.”

The system may need to understand:

• Product purpose
• Comfort
• Price
• Customer reviews
• Durability
• Availability
• Brand credibility
• User context

The brand that ranks highest for the phrase “walking shoes” is not automatically the product that best fits this request.

This creates a new visibility journey:

Rankability → Discoverability → Understandability → Trust → Recommendation

Stage 1: Rankability

The first stage is traditional search visibility.

Can search engines find and rank the brand’s content?

This still matters.

A company with a technically weak website, poor content or unclear product information may struggle in both traditional and AI-driven discovery.

SEO remains part of the foundation.

Stage 2: Discoverability

Being rankable doesn’t automatically mean AI systems can easily find all the relevant information.

The brand needs a strong digital footprint across sources such as:

• Website
• Product pages
• Merchant feeds
• Marketplaces
• Reviews
• News mentions
• Industry sites
• Authoritative directories

The goal is to create enough reliable signals for AI systems to come across the brand in relevant situations.

Stage 3: Understandability

Being found is not enough.

AI needs to understand:

• What does this company sell?
• Who are its products for?
• What problems do they solve?
• How are they different?

Imagine two products.

Product A:

“Advanced performance technology for modern users.”

Product B:

“Lightweight waterproof hiking jacket designed for day hikes and changing weather.”

The second description gives far more useful context.

AI visibility depends heavily on clarity.

Stage 4: Trust

Before recommendation comes trust.

AI systems may come across many different claims about a product.

The brand’s own website is only one source.

Other signals can include:

• Reviews
• Independent articles
• Retail listings
• Industry publications
• Expert mentions
• Customer discussions
• Authoritative third-party sources

This is why digital PR and reputation are increasingly linked with AI visibility.

A company can’t build strong recommendation signals entirely through claims it publishes about itself.

Independent confirmation matters.

Stage 5: Recommendation

Recommendation is the most valuable stage.

At this point, an AI system has enough relevant information to bring up the product when the customer’s need matches it.

For example:

“What are some reliable carry-on suitcases under $150 for frequent travel?”

A brand may become part of the answer because the system understands:

• Its product category
• Price range
• Product attributes
• Customer reputation
• Merchant availability

This is the start of the AI-Recommended Brand.

Recommendation Visibility Could Become a New Competitive Metric

Brands already track:

• Google rankings
• Search impressions
• Click-through rates
• Marketplace rankings
• Advertising visibility

AI brings in possible new metrics:

• How often is the brand mentioned?
• Which prompts trigger it?
• Which competitors appear instead?
• How is the brand described?
• Is the information accurate?
• Which sources support the recommendation?

This creates a new form of competitive intelligence.

Instead of only tracking:

“Who ranks above us?”

brands can begin asking:

“Who does AI recommend instead of us, and why might that be happening?”

What Walmart’s Strategy Means for Ecommerce Brands

Most ecommerce companies don’t have Walmart’s budget, stores or technology.

But the strategic lessons from Walmart’s approach are still useful.

The main lesson is not:

“Every brand should build what Walmart built.”

The better lesson is:

“Every brand should prepare its product information for a world where AI increasingly shapes discovery.”

Several practical actions follow from this.

1. Don’t Depend Only on Keyword Rankings

Keyword rankings are still valuable.

But customer discovery is getting more scattered.

People can now find products through:

• Google
• Marketplaces
• Social platforms
• ChatGPT
• Gemini
• AI search
• AI shopping interfaces

An ecommerce strategy built entirely on traditional rankings may leave gaps in visibility.

Brands should keep investing in SEO while also testing how their products appear in AI-driven experiences.

2. Make Product Information Machine-Readable

Good product information should work for both humans and machines.

Brands should clearly define:

• Product name
• Brand
• Price
• Availability
• Category
• Attributes
• Variants
• Product identifiers
• Images

This can be supported through structured data and accurate feeds.

The goal is to remove uncertainty.

3. Strengthen Brand and Entity Authority

AI systems need to understand who a company is.

Brand information should be consistent across:

• Website
• Social profiles
• Product feeds
• Marketplaces
• Business listings
• Publications
• Third-party references

A brand that describes itself differently in every place creates needless confusion.

Consistency helps strengthen entity understanding.

4. Maintain Accurate Product Feeds

Product feeds shouldn’t be treated as a one-time setup.

Prices, inventory and products change constantly.

Feeds need regular updates.

OpenAI’s product-feed guidance explicitly recommends keeping price and availability current so ChatGPT can give accurate product discovery information.

For ecommerce businesses, feed quality may increasingly become part of AI visibility performance.

5. Build Trusted Third-Party References

A brand saying:

“Our product is the best.”

is just marketing.

When independent sources consistently speak well of the product, that creates stronger outside context.

Brands can strengthen their digital footprint through:

• Digital PR
• Product reviews
• Industry publications
• Genuine customer feedback
• Expert features
• Useful partnerships
• Relevant directories

The goal is not to create fake mentions.

It is to build a credible reputation across the web.

6. Improve Reviews and Reputation

Reviews give useful real-world information about products.

They can show whether customers consistently value:

• Comfort
• Quality
• Reliability
• Ease of use
• Customer service

Strong reputation management therefore matters beyond improving conversions.

It may also strengthen the wider information environment around a product.

7. Answer Real Customer Questions

Brands should create content around the questions customers really ask before buying.

For example:

Instead of publishing only:

“Running Shoes Product Page”

also answer:

“Which running shoes are best for pavement?”

“How much cushioning does a beginner need?”

“What should I look for if I run three times per week?”

This kind of content gives context that answer engines and AI systems can read and use.

8. Prepare for Conversational Product Discovery

Ecommerce businesses should test their products with natural-language prompts.

Instead of only checking rankings for:

“best blender”

try:

“I make frozen fruit smoothies every morning and need an easy-to-clean blender under $150.”

Does the brand appear?

If not, which competitors do?

What information do those competitors provide that your brand doesn’t?

This kind of analysis can reveal AI visibility gaps.

What Smaller Retailers Can Learn From a Fortune 500 Company

A smaller ecommerce company can’t copy Walmart’s infrastructure.

Walmart has:

• Huge inventory
• Large customer datasets
• Thousands of physical stores
• Massive delivery capacity
• Marketplace sellers
• Large technology teams

A smaller business may have none of these.

That doesn’t mean it can’t prepare for AI commerce.

In fact, many of the improvements needed for AI visibility can already be done with normal ecommerce and content systems.

Improve Product Content First

Start with the basics.

Every important product should clearly explain:

• What it is
• Who it’s for
• Why it’s useful
• Important features
• Specifications
• Dimensions
• Compatibility
• Materials
• Use cases

Avoid descriptions full of vague marketing language.

Clarity is more useful than hype.

Use Structured Data

Schema markup can help search engines and other systems understand the key information on a page.

For ecommerce sites, useful structured data can cover:

• Products
• Offers
• Reviews
• FAQs
• Organisations
• Breadcrumbs

Schema doesn’t guarantee AI recommendations.

But it supports clearer machine-readable information.

Keep Merchant Feeds Clean

If the business already uses product feeds for advertising or marketplaces, review them regularly.

Check for:

• Missing titles
• Weak descriptions
• Wrong prices
• Out-of-stock items
• Missing identifiers
• Broken images
• Incorrect variants

The feed should represent the product accurately.

Build Useful FAQs

Frequently asked questions can help answer real buying concerns.

For example:

“Is this waterproof?”

“Does it work with an iPhone?”

“What size should I choose?”

“Can I wash it in a machine?”

These questions often hold exactly the kind of context that shoppers later ask AI assistants.

Create Comparison Content

Customers regularly compare products before buying.

Brands can help by creating honest comparisons.

For example:

• Model A vs Model B
• Beginner option vs premium option
• Small home vs large home

Good comparison content should explain the differences clearly instead of pretending one product is always better.

Explain Product Use Cases

Many product pages explain features but not situations.

Instead of only saying:

“500W motor.”

explain:

“Suitable for daily smoothies using frozen fruit and ice.”

Use cases connect technical specifications with customer needs.

That can make products easier for both people and AI systems to understand.

Build Brand Mentions

Smaller brands should also invest in reputation outside their own website.

Possible sources include:

• Niche publications
• Relevant blogs
• Industry websites
• Podcasts
• Expert interviews
• Genuine product reviews
• Local or industry media

This improves brand discovery and authority.

Invest in Digital PR

Digital PR can support:

• Brand awareness
• SEO
• Referral traffic
• Authority
• AI visibility

The strongest campaigns usually offer something genuinely useful:

• Research
• Data
• Expert insight
• Industry analysis
• Original stories

Instead of paying for low-quality mentions, businesses should aim to earn credible coverage.

Monitor AI Visibility

Smaller brands may actually have an advantage here.

They can start testing quickly.

A business can make a list of 20–50 commercially important prompts and track them across AI platforms.

The goal is to understand:

• Where do we appear?
• Where don’t we appear?
• Which competitors appear again and again?
• What information about us is wrong?

AI visibility monitoring doesn’t need Walmart-sized technology.

It needs a consistent way of measuring.

A Practical AI Visibility Framework for Ecommerce Brands

To make AI visibility easier to understand, ecommerce brands can use a simple framework.

We can call it the:

D-U-T-R Framework

• Discoverable
• Understandable
• Trustworthy
• Recommendable

The framework helps show why a brand may or may not appear in AI-driven product discovery.

D — Discoverable

The first question is:

Can AI systems find the brand and product?

A product can’t be recommended if there’s little reliable information about it.

Discoverability can come from:

• Indexed website pages
• Product pages
• Merchant feeds
• Marketplaces
• Reviews
• Publications
• Business profiles
• Industry websites

Discoverability Questions

Ask:

• Does the product have its own indexable page?
• Can search engines access it?
• Is it included in relevant product feeds?
• Is the brand mentioned outside its own website?
• Are important products covered by third-party sources?

If the answer is mostly no, the brand may have a discoverability problem.

U — Understandable

Next, ask:

Can AI correctly understand the product?

Being visible is not enough.

The system needs to understand:

• Product type
• Purpose
• Audience
• Features
• Price
• Use cases
• Variants
• Difference from alternatives

Understandability Questions

Ask:

• Is the product title clear?
• Does the description explain real use cases?
• Are important attributes available?
• Is structured data set up correctly?
• Are brand and product names consistent?
• Can someone understand the product without already knowing the brand?

If the information is vague, AI systems may struggle to match the product to detailed customer requests.

T — Trustworthy

The third stage asks:

Is there enough evidence to trust the information?

Trust can be supported by:

• Genuine customer reviews
• Independent mentions
• Industry coverage
• Consistent product information
• Clear business details
• Good customer service
• Authoritative sources

Trust Questions

Ask:

• Do independent websites mention the brand?
• Are there genuine customer reviews?
• Are product claims backed up?
• Is company information consistent?
• Is pricing reliable?
• Do users report positive experiences?

Trust is especially important because AI systems often work with information from multiple sources instead of relying only on a company’s own claims.

R — Recommendable

The final question is:

Does AI have enough reliable context to recommend this product for a relevant need?

This is where all the earlier stages come together.

A recommendable product is:

• Discoverable
• Clearly understood
• Backed by trustworthy information
• Relevant to the customer’s request

For example:

A customer asks:

“I need a lightweight waterproof jacket for weekend hiking under $150.”

For a product to become relevant, AI may need to understand that:

• It is a jacket
• It is waterproof
• It is lightweight
• It is designed for hiking
• It falls within the budget
• It is available
• Customers generally view it positively

This is why recommendation visibility can’t be solved with one SEO tactic.

It depends on an ecosystem of information.

How Brands Can Measure Their AI Shopping Visibility

AI visibility shouldn’t be treated as a vague branding idea.

It can be tested.

The exact analytics available from AI platforms are still developing, but brands can already build practical monitoring systems.

The starting point is to measure how often and how accurately the business appears when people ask commercially relevant questions.

Build an AI Shopping Prompt Set

Start by making a list of prompts that reflect real customer needs.

Don’t test only your brand name.

If someone already searches for your exact brand, they already know you.

The more useful test is:

Does the brand appear when the customer has a problem but hasn’t chosen a brand yet?

Useful prompt formats include:

Best X for Y

“Best running shoes for beginners.”

“Best laptop for university students.”

Compare X vs Y

“Compare Brand A and Brand B for home use.”

What Should I Buy For…?

“What should I buy for a small home office?”

Affordable Alternatives

“Affordable alternatives to Product X.”

Best Brands

“Best brands for lightweight travel luggage.”

Constraint-Based Prompts

“Best office chair under $300 for someone working eight hours per day.”

These prompts show far more about AI visibility than branded searches.

Test Across Multiple AI Environments

Don’t monitor only one platform.

Customer journeys are increasingly spread across several AI experiences.

A useful monitoring set can include:

• ChatGPT
• Gemini
• Google AI Search experiences
• Perplexity
• Other relevant AI discovery tools

This is especially important because each system uses different data sources, models and interfaces.

For example, Google has increasingly connected Gemini models with its Shopping Graph, which holds tens of billions of product listings. Google said in 2026 that billions of those listings are refreshed every hour.

Perplexity also supports conversational product research and merchant checkout experiences in supported markets.

So the goal is not to optimise for one AI answer.

It is to understand the brand’s wider AI discovery footprint.

Track Brand Mentions

The simplest metric is:

Was the brand mentioned?

Record:

• Yes
• No
• Position within the response
• Context of the mention

Repeat the same test over time.

A brand appearing once means less than a brand appearing consistently across many relevant prompts.

Track Recommendation Frequency

A stronger metric is recommendation frequency.

For example:

Suppose a company tests 50 prompts.

The brand appears in 18.

AI Recommendation Visibility:

18 ÷ 50 = 36%

This doesn’t mean the business has a 36% market share.

It is simply an internal visibility metric that helps track change.

The company can repeat the same prompt set every month and watch whether visibility is improving.

Track Citation and Source Visibility

Another important question is:

Which sources appear to support the answer?

Look for:

• Your website
• Product pages
• Review websites
• News articles
• Marketplaces
• Industry publications
• Competitor pages

This can reveal which outside sources matter in your information ecosystem.

For example, if AI regularly cites a strong independent review site, earning credible coverage there may be more valuable than producing another generic blog post.

Track Competitor Presence

Don’t record only your own brand.

Track competitors too.

For each prompt, note:

• Which competitors appear?
• Which appear most often?
• Which are mentioned first?
• What reasons are given?
• Which sources support them?

This creates an AI Share of Voice view.

For example:

• Brand A: 42% visibility
• Brand B: 37%
• Your Brand: 21%

Again, this should be treated as an internal research metric and not a universal platform metric.

Its value comes from tracking the same method consistently.

Track Product Accuracy

Visibility isn’t valuable if the information is wrong.

Check whether AI gives accurate:

• Product names
• Prices
• Features
• Availability
• Brand information
• Product descriptions
• Use cases

A wrong recommendation can sometimes be worse than no recommendation.

For example:

If AI says a product supports a feature that it doesn’t actually support, customers may lose trust.

Accuracy should therefore be part of AI visibility monitoring.

Track Recommendation Context

A brand may appear often but for the wrong reason.

For example:

A premium product might show up mainly in:

“cheapest alternatives”

when the brand actually wants to be known for quality.

Track the context.

Ask:

How is AI positioning us?

Possible positioning could include:

• Affordable
• Premium
• Reliable
• Beginner-friendly
• Professional
• Sustainable
• High-performance

This helps brands understand not only whether they appear but how AI understands them.

Create an AI Visibility Scorecard

Brands can combine these signals into a simple scorecard.

Metric	What to Measure
Brand Mention Rate	How often the brand appears
Recommendation Rate	How often it is actively recommended
Competitor Share	How often competitors appear
Source Visibility	Which sources support recommendations
Product Accuracy	Whether product information is correct
Context Accuracy	Whether positioning matches the brand
Platform Coverage	Visibility across different AI environments

This gives a repeatable way to monitor AI discovery.

Measure Trends, Not One-Off Answers

AI outputs can change.

A single prompt on a single day shouldn’t be treated as final.

A better process is:

Create fixed prompt set ↓ Test across platforms ↓ Record results ↓ Repeat regularly ↓ Compare change

For example:

• Month 1: 18% recommendation visibility
• Month 2: 23%
• Month 3: 31%

This gives the company a directional signal.

The same process can also reveal falling visibility.

Connect AI Visibility With Business Outcomes

Eventually, the most useful measurement will link AI discovery with commercial results.

Possible indicators include:

• Referral traffic from AI platforms
• AI-driven product page visits
• Assisted conversions
• Brand search growth
• Direct traffic
• Conversion rate
• New customer acquisition

OpenAI is also building structured commerce infrastructure around merchant feeds and product discovery, which should make it easier over time to study the link between AI discovery and transactions.

The Bigger Measurement Question

Traditional SEO taught businesses to track:

Where do we rank?

AI visibility calls for a wider question:

When customers describe their needs to AI, how often does our brand become part of the answer?

That is likely to become one of the most important new visibility questions for ecommerce marketers.

Brands that start measuring it now can set a baseline before AI-assisted product discovery becomes even more deeply built into the shopping journey.

Risks and Limitations of AI-Driven Product Discovery

AI-powered shopping can make product discovery faster and more convenient, but it isn’t perfect.

Any serious case study about conversational commerce should also look at the risks.

AI can help customers research, compare and discover products, but the quality of the experience depends heavily on the quality of the information the system has.

This creates several important limitations.

Incorrect Information Can Damage the Shopping Experience

AI systems can sometimes present information that is incomplete, unclear or wrong.

A product may be described with the wrong feature.

A compatibility detail may be misunderstood.

A comparison may oversimplify an important difference.

This is a bigger problem in commerce than in general information.

If AI describes a product incorrectly, the customer may spend money based on that mistake.

For example:

A shopper asks for a laptop that supports a specific feature.

The AI recommends a model.

The customer later finds out that the model doesn’t actually support it.

This can lead to:

• Returns
• Customer frustration
• Lost trust
• Poor reviews
• Support costs

That is why product accuracy matters so much in AI commerce.

Brands can’t treat product information as basic marketing copy.

It needs to be clear, current and factual.

Pricing Can Change Faster Than AI Information

Ecommerce prices can change often.

A retailer may update a price because of:

• Promotions
• Demand
• Inventory
• Seasonal sales
• Seller changes
• Currency movements

If AI shows an old price, the experience quickly becomes less useful.

For example:

• AI result: $79
• Retailer page: $99

Even if the gap is caused by a normal price update, the customer may feel the recommendation was misleading.

This is one reason fresh merchant feeds are important.

AI shopping works best when pricing data is updated often.

The closer product discovery moves to checkout, the more important real-time accuracy becomes.

Inventory Can Change Quickly

Availability creates a similar problem.

A product may be in stock when an AI system finds it and out of stock by the time the customer decides to buy.

This is especially common with:

• Popular products
• Limited stock
• Seasonal items
• Marketplace sellers
• Local inventory

Imagine a customer spending time comparing products through AI, choosing one, and then finding out it can’t be bought.

That creates friction instead of reducing it.

For agentic commerce to work well, AI systems need stronger and stronger links with live inventory and merchant systems.

Product discovery can’t be treated separately from availability.

Recommendation Transparency Remains Important

Another major question is:

Why was this product recommended?

Traditional search engines already raise questions about ranking.

AI recommendations can be even harder for customers to make sense of.

A user may see three products and wonder:

• Why these three?
• Why not another brand?
• Was price the main factor?
• Did reviews matter?
• Was the merchant chosen because of availability?
• Is the recommendation sponsored?

A clear separation between organic recommendations and advertising is therefore important.

Consumers need to know when they’re seeing a recommendation based on relevance and when they’re seeing paid promotion.

The more AI influences buying decisions, the more important recommendation transparency becomes.

Merchant Data Quality Can Become a Competitive Issue

AI systems depend heavily on merchant information.

If the merchant gives poor data, the AI may misunderstand the product.

Common problems can include:

• Weak product titles
• Missing attributes
• Wrong prices
• Poor images
• Duplicate products
• Missing variants
• Outdated stock information
• Confusing brand names

This holds an important lesson for ecommerce businesses.

AI visibility isn’t only a marketing problem.

It is also a data-quality problem.

A brand can have excellent advertising and still struggle in AI shopping if its catalogue data is unreliable.

Consumer Trust Cannot Be Assumed

People may use AI for product research, but that doesn’t mean they’ll automatically trust every recommendation.

For expensive or high-risk purchases, customers may still want to:

• Read independent reviews
• Visit the retailer’s website
• Compare other stores
• Watch product videos
• Check expert opinions
• Ask friends or family

This means conversational commerce is more likely to support the decision journey than to fully replace all other research.

Trust will depend on:

• Recommendation quality
• Information accuracy
• Clear sources
• Transparent pricing
• Reliable merchants

If AI repeatedly gives poor recommendations, customers may stop using it for shopping.

Personalisation Creates Privacy Questions

Personalisation can make shopping more useful.

A system that understands preferences, budget, location and past purchases can give better recommendations.

But more personalisation also raises privacy questions.

Consumers may ask:

• What information is being used?
• Is purchase history involved?
• Is location being used?
• Are preferences stored?
• Can users control personalisation?

The value exchange has to stay clear.

Customers may accept personalisation when it clearly improves the experience.

They may become uncomfortable if the system seems to know more than expected.

Retailers and AI platforms therefore need to balance convenience with user control and privacy.

Brands May Become Too Dependent on AI Platforms

Platform dependence is another strategic risk.

Many ecommerce businesses already depend heavily on:

• Google
• Amazon
• Meta
• Marketplaces
• Retail media platforms

AI could add another layer of dependence.

If a large share of discovery starts inside a small number of AI platforms, changes to those systems could influence which brands get visibility.

A business should therefore avoid building its whole strategy around one AI platform.

The safer approach is diversified visibility across:

Search + Marketplaces + Direct Traffic + Social + Email + AI Discovery

AI should become another channel.

It shouldn’t become the only channel.

AI Commerce Still Needs Human Oversight

AI can automate parts of shopping, but it shouldn’t remove human judgement completely.

Customers still need control over:

• Product choice
• Budget
• Payment
• Delivery
• The final purchase decision

Merchants also need oversight of:

• Product data
• Pricing
• Inventory
• Customer service
• Returns

The best use of AI may therefore be assistance and not total automation.

A good commerce agent should cut unnecessary work while keeping the customer in control.

Is Traditional Product Search Really Being Replaced?

The title of this case study asks whether Walmart is replacing traditional product search with AI.

The most accurate answer is:

Not completely.

Traditional product search is not disappearing overnight.

Search bars, filters, category pages and product listings are still useful.

For simple searches, they may still be the fastest option.

If a customer already knows exactly what they want, there may be no reason to have a long conversation.

For example:

“Nike Air Max size 9.”

or:

“iPhone 17 case.”

These queries are direct.

Traditional search handles them well.

The real change happens when the customer’s need is more complicated.

Search Works Best When the Customer Knows the Product

Traditional search is strongest when the user knows the:

• Product name
• Brand
• Model
• Category
• Specification

The customer gives the system a clear instruction.

The search engine returns matching results.

This model is still efficient.

AI Becomes More Useful When the Customer Knows the Problem

Conversational discovery becomes more useful when the customer says:

“I need something to keep my bedroom cooler at night without using too much electricity.”

Now the user is describing a problem and not a product.

The system may need to consider:

• Fan
• Air cooler
• Portable AC
• Room size
• Electricity usage
• Budget
• Noise

This takes more reasoning than a simple keyword match.

That is where AI can add value.

Traditional Search and AI Will Likely Coexist

The more realistic future is not:

AI replaces search.

It is:

AI changes when and how people use search.

A customer may:

1. Ask AI what type of product they need.
2. Compare several options.
3. Visit Walmart.
4. Use filters.
5. Read reviews.
6. Complete the purchase.

Another customer may complete more of the journey inside an AI interface.

Different shoppers will behave differently.

Different products will also need different journeys.

Keyword Search Is Becoming One Part of a Wider Discovery System

The stronger conclusion is:

Keyword search is increasingly being supplemented by conversational and agent-driven discovery.

That difference matters.

Traditional ecommerce search is built around:

“What did the customer type?”

AI commerce increasingly asks:

“What is the customer trying to accomplish?”

Both models can exist together.

The difference is that product discovery is becoming more flexible.

Retailers are moving from one main discovery interface toward several:

• Search bars
• Filters
• Recommendations
• AI chat
• Visual shopping
• Voice
• Shopping agents

Walmart’s strategy reflects this wider shift.

The company isn’t simply removing its search bar.

It is creating another way for customers to reach the right product.

The Future: From Search Engines to Shopping Agents

Digital product discovery has changed several times.

The next change appears to be a move toward AI systems that do more than answer questions.

A useful way to understand the evolution is:

Search Engines ↓ Answer Engines ↓ Recommendation Engines ↓ AI Shopping Assistants ↓ Commerce Agents

This doesn’t mean every platform will follow exactly this path.

It shows the direction the industry seems to be exploring.

Stage 1: Search Engines

Search engines help people find information.

The customer asks:

“best blender for smoothies”

The search engine gives links.

The user does most of the research.

Stage 2: Answer Engines

Answer engines reduce the need to open every result.

The customer asks:

“What should I look for in a smoothie blender?”

The system explains the important features.

The customer gets an answer and not just a list of pages.

Stage 3: Recommendation Engines

The next stage is more personal.

The customer says:

“I make frozen fruit smoothies every morning, need something easy to clean and want to spend under $150.”

The system can recommend products that match those needs.

Now AI is influencing which products get considered.

Stage 4: AI Shopping Assistants

Shopping assistants can help manage more of the journey.

They may help with:

• Research
• Comparison
• Budget
• Features
• Availability
• Reviews
• Product selection

The customer stays in the conversation while refining the decision.

This is already becoming visible in modern AI shopping experiences.

Stage 5: Commerce Agents

The final stage is more agentic.

The customer states a goal.

For example:

“Help me buy everything I need for a weekend camping trip for two people under $300.”

A commerce agent could potentially help to:

1. Understand the task.
2. Build the shopping list.
3. Compare options.
4. Stay within budget.
5. Check availability.
6. Suggest alternatives.
7. Prepare the purchase.
8. Ask the customer for final approval.

The system moves from answering questions to helping complete a commercial task.

The Customer May Move From Searching to Delegating

This may be the biggest change in behaviour.

Traditional search asks the customer to do most of the work.

The future may involve more delegation.

Instead of:

“Show me products.”

customers may increasingly say:

“Help me solve this problem.”

That changes the role of ecommerce technology.

The system becomes less like a catalogue and more like an assistant.

Retailers Will Still Need Strong Infrastructure

Even advanced commerce agents can’t replace good retail operations.

They still need accurate:

• Product data
• Inventory
• Pricing
• Payments
• Delivery
• Returns
• Customer support

This is why Walmart is such an interesting example.

The AI layer can become more powerful because it sits on top of a large commerce infrastructure.

The future of AI commerce is therefore unlikely to be controlled by AI alone.

It will depend on the connection between:

Intelligence + Data + Retail Infrastructure + Customer Trust

This Is an Emerging Direction, Not a Finished Model

It’s important not to treat this future as guaranteed.

AI shopping is still developing.

Customer behaviour may change slowly.

Some product categories may adopt conversational shopping faster than others.

Technical, regulatory and trust issues will also affect adoption.

Still, the direction is getting clearer.

Product discovery is gradually moving from:

Find information

to

Understand options

to

Recommend products

to

Help complete the purchase

That is the wider context behind Walmart’s investment in ChatGPT and agentic commerce.

Key Takeaways From the Walmart + ChatGPT Case Study

The Walmart and OpenAI case study offers several important lessons for ecommerce brands.

1. Shopping Discovery Is Becoming More Conversational

Customers no longer need to communicate only through keywords.

They can describe:

• Goals
• Problems
• Budgets
• Preferences
• Use cases

This lets AI systems understand shopping intent in more detail.

2. Walmart Is Treating AI as a Commerce Interface

The strategy goes beyond adding a chatbot.

AI is becoming another way into Walmart’s retail ecosystem.

The customer journey can increasingly move from:

Conversation → Discovery → Comparison → Commerce

3. Product Data Is Becoming Strategically Important

AI systems need reliable product information.

Accurate:

• Titles
• Descriptions
• Prices
• Availability
• Variants
• Attributes

can improve the quality of product discovery.

Product feeds may therefore become more and more important to AI commerce.

4. AI Visibility Is Becoming a Marketing Consideration

Brands have traditionally focused on:

Google visibility

and

Marketplace visibility

They may increasingly need to monitor:

AI recommendation visibility

This creates a new question:

Does AI understand and bring up our brand for relevant customer needs?

5. Search and Commerce Are Moving Closer Together

The traditional journey often keeps research and purchase apart.

AI shopping can connect:

Research + Comparison + Recommendation + Commerce Action

This can shorten the path from interest to purchase.

6. Fulfilment Still Matters After AI Discovery

A great recommendation is not enough.

The product still needs to be:

• Available
• Correctly priced
• Deliverable
• Easy to buy

Walmart’s physical stores, inventory and delivery network remain major strengths.

AI improves discovery.

Retail infrastructure completes the experience.

7. Brands Need Visibility Beyond Traditional SERPs

Search engine rankings are still important.

But brands now need to think about a wider ecosystem:

• Search Engines
• Marketplaces
• AI Answers
• AI Recommendations
• AI Shopping Experiences

The goal is no longer just to rank.

It is to be:

Discoverable, understandable, trustworthy and recommendable.

Final Conclusion: Will AI Recommend Your Products?

Walmart’s move toward ChatGPT and agentic commerce shows that ecommerce discovery is beginning to change.

The old model was built mainly around search.

The customer typed keywords.

The retailer returned products.

The customer did the research.

The customer compared the options.

The customer completed the purchase.

AI brings in a different model.

The customer can describe a need.

The system can help understand that need.

It can research products.

It can compare options.

It can narrow the choices.

And more and more, it can connect the customer directly with commerce.

This doesn’t mean traditional ecommerce is going away.

Search bars, product pages, marketplaces, SEO and paid advertising will continue to matter.

But another discovery layer is growing next to them.

That layer is AI recommendation.

The New Ecommerce Competition

For years, ecommerce brands competed around questions such as:

• Who ranks first on Google?
• Who appears highest in Shopping ads?
• Who ranks better on Amazon or Walmart?

Those questions are still important.

But another question is emerging:

When a customer asks AI what they should buy, which brands become part of the answer?

That question changes the optimisation challenge.

A brand needs more than keywords.

It needs clear:

• Product information
• Brand identity
• Reviews
• Structured data
• Merchant feeds
• Pricing
• Availability
• Trusted references
• Useful, customer-focused content

AI systems need enough context to understand not only what the product is, but when it is relevant.

From Ranking First to Being Recommended

The next ecommerce competition may not only be about who ranks first on Google.

It may also be about which brands AI systems can:

Find ↓ Understand ↓ Trust ↓ Recommend

at the exact moment a customer asks:

“What should I buy?”

That is a fundamentally different type of visibility.

It is the difference between showing up in a list and becoming part of a recommendation.

Is Your Brand Ready for AI-Driven Discovery?

Ecommerce businesses should start asking practical questions now:

• Can ChatGPT understand our products?
• Does Gemini recognise our brand for the right topics?
• Does our website clearly explain product use cases?
• Are our merchant feeds accurate?
• Do AI systems mention competitors more often than us?
• Is our product information consistent across the web?
• Are trusted third-party sources talking about our brand?

These questions can help show whether a business is ready for a more AI-driven customer journey.