For years, retailers have fought for the top spot on Google. Now, they face a new challenge: persuading artificial intelligence that their product is the one worth recommending.
Consumers are increasingly using tools such as ChatGPT, Gemini and AI-powered search to research products, compare features and narrow down their options. Instead of scrolling through pages of links, shoppers can ask a detailed question and receive a shortlist within seconds.
That changes the rules of product discovery. Being searchable is no longer enough. Retailers must ensure their products can be accurately understood, compared and recommended by machines.
AI is beginning to sit between the consumer and the retailer at some of the most influential moments in the buying journey.
A shopper might ask for “the best lightweight pushchair for city travel under £400” or “a sustainable party dress that can arrive by Friday”. The AI assistant can interpret those requirements, compare the available options and present a handful of recommendations.
The customer may never see the dozens of brands that were considered and excluded along the way.
Google has been steadily embedding this behaviour into its existing search ecosystem. According to an update published by Google in May 2025, AI Overviews had expanded to more than 200 countries and territories and over 40 languages. Google also reported that, in major markets including the US and India, AI Overviews were driving a more than 10% increase in usage for the types of queries where they appeared.
Its shopping capabilities have since moved further into conversational discovery. In April 2026, Google said its Shopping Graph contained more than 50 billion product listings, with two billion updated every hour. This information now powers product suggestions, comparisons, prices and availability within Gemini and AI Mode.
ChatGPT is taking a similar path. In March 2026, OpenAI introduced richer product-discovery tools that allow users to browse products visually and compare prices, reviews and features without repeatedly moving between websites.
OpenAI also announced that retailers including Sephora, Target, Best Buy, Nordstrom and Wayfair had integrated with its Agentic Commerce Protocol for product discovery. Shopify’s product catalogue was also connected to ChatGPT, giving millions of merchants a route into its shopping results.
The direction of travel is clear: AI platforms do not simply want to answer shopping questions. They want to become part of the shopping interface.
It is important not to overstate the shift. Traditional search, marketplaces, social platforms and retailers’ own channels still account for most online product discovery. AI referrals are growing from a comparatively small base, and the platforms themselves remain in rapid development.
Yet the behaviour of shoppers arriving through AI deserves attention.
Adobe Analytics reported in January 2026 that traffic from generative AI tools to US retail websites increased by 693% year on year during the 2025 holiday period. More significantly, those visits converted 31% better than traffic from other sources.
Adobe’s analysis, which covered more than one trillion visits to US retail websites, also found that AI-referred shoppers spent 45% longer on retail sites, viewed 13% more pages and were 33% less likely to leave immediately.
That suggests AI is not necessarily creating another stream of casual browsers. By the time these consumers reach a retailer, they may have already explained their needs, compared alternatives and moved considerably closer to a decision.
For retailers, the commercial question is therefore bigger than traffic volume. It is about influence: which brands are being introduced during the consideration phase, and which are missing from the conversation entirely?
Success in this environment will not come from simply publishing more content.
AI assistants need clear, consistent and current information. Product titles, descriptions, specifications, images, prices, stock availability, delivery details, returns policies, reviews and FAQs can all affect whether a product is surfaced—and whether the recommendation is accurate.
This presents a familiar problem for large retailers. Information is often distributed across ecommerce platforms, product information management systems, marketplaces, local websites, customer-service tools, PDFs and campaign landing pages.
A discontinued product may still appear in an old buying guide. A delivery policy might differ between a webpage and a downloadable document. Materials, measurements or compatibility details may be missing from a product feed, even though they appear elsewhere on the website.
To a customer, these may appear to be small inconsistencies. To an AI system trying to provide one confident answer, they create uncertainty.
The retailers best prepared for AI discovery will not necessarily be those producing the most content. They will be those with the strongest control over their product data, and the clearest version of the truth.
There is also a risk in treating AI optimisation as simply the latest version of SEO.
Retailers must consider not only whether an AI platform can find a product, but how it represents the brand. A recommendation may prioritise price and specifications while overlooking customer service, quality, loyalty benefits or the wider experience surrounding the purchase.
There are commercial questions too. If more discovery and comparison happen inside third-party AI interfaces, retailers could lose valuable website visits, customer insight and control over the relationship.
As advertising, sponsored offers and native checkout become more common within AI platforms, the distinction between an independent recommendation and a paid placement will also require close scrutiny.
In January 2026, Google announced that it was testing “Direct Offers” within AI Mode, allowing advertisers to surface exclusive discounts when shoppers were considered ready to buy. It also outlined checkout capabilities that would allow eligible shoppers to buy from retailers through Google’s AI surfaces, while the retailer remained the seller of record.
AI visibility, therefore, should not be pursued at any cost. Retailers need to decide which information they are willing to distribute, which platforms deserve access and how they will protect their customer relationships.
The immediate priority is not to chase every new AI platform. It is to strengthen the foundations that make products discoverable wherever shoppers choose to look.
Retailers should:
AI is unlikely to replace every existing shopping channel. But it is becoming a powerful new layer between a consumer’s question and a retailer’s answer.
The brands at greatest risk are not simply those that fail to appear. They are those that allow fragmented systems, incomplete product information or third-party platforms to tell their story for them.
The next battle for retail visibility has already begun, and much of it is taking place before the customer ever reaches the website.