AI search shift leaves products invisible to shoppers

Akeneo has published research showing that more than two-thirds of Google searches end without a click to an external website. The share rises to about 80% when AI-generated summaries appear.

The findings point to a shift in how consumers find products online, as AI assistants increasingly surface recommendations directly instead of sending shoppers to retailer sites.

For brands and retailers, that creates a different kind of competition in digital discovery. Rather than relying mainly on traditional search rankings to attract traffic, companies must ensure product information can be read and interpreted by AI systems.

The report argues that incomplete or poorly structured product data can leave items effectively hidden from AI-generated answers. If specifications are missing, buried in PDF files, or presented as unstructured text, AI systems may struggle to identify a product accurately enough to recommend it.

Research cited in the study found hallucination rates across leading AI models range from 15% to 52% when structured product data is absent. In those cases, AI tools may infer missing details or steer users towards rival products with clearer information.

Search shift

The report also highlights what it describes as a weakening link between conventional search engine rankings and visibility in AI-generated responses. It says 47% of AI Overview citations now come from pages ranking below position five in standard search results.

That suggests machine-readable product content is becoming a separate route to discovery, even for brands that do not hold the top positions in search listings. For retailers, the practical question is not only whether a product page ranks well, but whether an AI system can verify and trust the information on it.

Akeneo frames this as a product-level contest rather than a site-level one. Generative AI systems increasingly answer specific consumer questions with individual product suggestions, meaning each buying decision can hinge on the quality of data attached to a single item.

The study links that change to a broader commercial trend, citing McKinsey estimates that AI-powered search could influence USD $750 billion of US consumer spending by 2028.

Invisible shelf

Akeneo uses the term "Invisible Shelf" to describe products that are available online but do not appear reliably in AI recommendations because their information is incomplete, insufficiently machine-readable, or not trusted. The report argues that marketing tools alone cannot solve the problem.

Instead, companies should treat AI discoverability as a product data issue. In practice, that means improving the structure, consistency, and governance of product information so AI systems can interpret it and match it to consumer intent.

One issue for retailers is that many product records were built for human browsing rather than machine interpretation. Descriptions may be written in free text, technical details may sit in attached documents, and data fields may differ across product ranges, making it harder for AI tools to compare items or confirm their attributes.

As AI-generated search summaries become more common, the cost of those weaknesses may grow. If a product is not recognised as a credible answer to a query, it may never enter the consumer's consideration set, even if it is competitively priced or well reviewed elsewhere.

Industry response

The report also points to the rise of answer engine optimisation, a set of methods aimed at tracking and improving how products appear in AI-generated responses. But it argues that lasting gains depend less on surface-level optimisation and more on the quality of the underlying product records.

That view reflects a wider question facing retailers as online discovery changes. If AI systems become a common starting point for purchase decisions, the quality of structured product information could become as important as search marketing has been over the past two decades.

The same logic extends beyond recommendation tools to what the report describes as agentic commerce, in which AI agents compare, select, and potentially buy products on behalf of consumers. In that setting, the data attached to a product would help determine whether an automated system considers it suitable at all.

The company behind the research says the warning is grounded in those broader shifts.

"The next battle for visibility won't be fought on search engine results pages," said Justin Thomas, Vice President of Sales, EMEA North, Akeneo. "Increasingly, products will be recommended directly by AI systems. If AI cannot understand, verify, and trust your product information, your products risk becoming invisible during one of the fastest-growing stages of the buying journey."