AI shopping assistants like ChatGPT and Gemini are re-writing the rules of product discovery. Instead of typing search terms into Google or going directly to a known retailer’s site, shoppers increasingly turn to AI to find products, compare options, and decide where to buy. Often, they’re doing it with unbranded queries, favoring solutions over brand recognition.

That shift has made AI recommendations an inescapable battleground for retail. Established retailers aren’t immune: Brands showing up in those recommendations are the ones whose product details AI agents can read, understand and recommend. AI-referred traffic to U.S. retail sites rose 393% year-over-year in Q1 2026, now converting 42% better than other traffic sources, according to Adobe. Brands that don't actively prepare for it risk becoming invisible, fast.

In the agentic commerce era, showing up in AI results runs through content and strategy, not the IT department. That makes it a problem retail leaders can own. The right commerce foundation makes that possible without a technical overhaul.

Why Most Product Pages Are Invisible to AI

AI agents don’t browse the way humans do. They scan for explicit, structured information — attributes, specs, materials, use cases, availability — and make recommendations based on what they can parse. Most product pages weren’t built for that. 

Adobe data shows the average AI readiness score for U.S. retail product pages is 66%, with the best-performing sites hitting 82.5% and the lowest at 54.2%. That gap reflects how unevenly retailers have adapted to new buying behaviors.

The fix has less to do with technology than context. “People don’t tend to search for technical details about products,” says Shaun McCran, Adobe’s product marketing lead for CXO Solutions. “They want an experience from a product. By creating the context for that experience, you’re able to match what they’re searching for with how the product is used.”

An overlooked barrier: AI crawlers can’t access product details buried in pop-ups, expandable sections, or page elements that only load after a user clicks. When LLMs can’t read that information, they can’t match shopper intent to your products.

Consider a coffee brand whose storefront reads only “medium roast, caramel flavor.” An AI agent has almost nothing to match against a query. Describe the same product as single-origin Colombian, USDA-certified organic, and ideal for pour-over brewing, and suddenly there’s context for the AI to work with — and a much stronger chance of being recommended. The same logic applies to technical details: a bean grown with a honey processing method leads to a specific flavor profile that shoppers are actively searching for, even if they’d never think to search for the altitude. Bridging that gap is where many brands fall short.

Product titles carry disproportionate weight in all of this. “Product titles and descriptions are the top values an LLM reads when it comes to your product pages,” says Alex Jose, senior product manager for Adobe Commerce. “When an LLM builds its knowledge of your products and brand, it gives the highest weight to those two areas.” Titles and descriptions built for typical commerce systems rarely carry the narrative product value that AI agents rely on to make confident recommendations.

Building Your AI Discoverability Strategy

When it comes to building AI discoverability, the most efficient place to start is with the products that matter most to revenue. Focusing first on hero SKUs and high-margin categories lets retailers see meaningful results without a full-catalog overhaul. Products receiving the most LLM traffic also deserve to be prioritized. “With Adobe Commerce, retailers know which product pages are being most crawled by LLMs, so they always have a data-driven starting point for where to focus first,” Jose explains. 

A basic readability audit is another practical way to pinpoint focus areas. Can an AI agent access your product’s key attributes, pricing, availability and return policy? Gaps in those areas weaken the odds of being recommended. Adobe’s AI Content Visibility Checker, a free tool, makes running that audit straightforward.

AI visibility is an ongoing discipline, not a project with a finish line. “It never stops, but it does get easier,” McCran says. “Once you’ve set the foundation, it enters a much more reasonable run-and-operate state. Think of it like SEO. LLM optimization is a similar commitment.” Without that discipline, BCG warns, retailers risk being reduced to background utilities in agent-controlled marketplaces, with shrinking direct access to customers. 

Win the Recommendation, Win the Sale

Shopper adoption of AI assistants is accelerating, and so are the tools available to them. Retailers who’ve invested in AI-readable product content are already reaping the benefits: AI-referred shoppers arrive more informed, more confident and ready to buy.

“There are over 500 different LLMs out there,” McCran says. “How do you adapt to this new shopping behavior and also drive greater value because of it?” 

McCran’s advice is to start small and move fast. “We always advocate test-and-learn. Try something, do it short and quick. See the results, reiterate, learn fast and reapply lessons. Don’t be afraid to fail, you’ll still learn something.” The same logic applies to scope: “Don’t run a 12-month project and find the results aren’t relevant anymore.”

AI discoverability rewards preparation. The product content decisions retailers make now will determine whose products get recommended — and whose get passed over.

▷ Learn how to drive greater product visibility, conversion, and loyalty across human and AI-led shopping.