Deloitte Digital published new B2C commerce research this week, and the headline is uncomfortable. Just about one in four brands believe they deliver an excellent online shopping experience. Far fewer actually do, according to the consumers Deloitte surveyed alongside them. The study paired 550 B2C company leaders with 1,000 recent shoppers and asked both what the experience is really like. The two groups described different worlds.
The gap is not a rounding error. Consumers were 20 times more likely than brands to say that commerce experiences are not keeping up with their expectations, and they reported that experiences have not meaningfully improved in the two years since Deloitte's last survey. This is happening while spending tightens: 57% of consumers said they spent less over the prior year, and 62% of brands said reduced spending hurt them. Every interaction matters more, and brands are grading themselves on a curve their customers do not recognize.
The confidence has a source. For two decades SEO was the scoreboard, and brands got good at it. Keyword rankings, page position, and organic traffic still look healthy, so leaders read the storefront as performing and stop investing in what they assume is handled. The catch is what that scoreboard measures: a shopper typing a query and scanning links. It says nothing about whether an AI shopping agent can read your product data and recommend you, and shoppers increasingly research, find, and buy through those agents.
Brands are acing last decade's exam and assuming it still counts.
The gaps show up in the basics: accurate search, clear availability, reliable delivery information, straightforward pricing, responsive service. These used to be differentiators. Now they are the price of admission, and most consumers say brands have not improved on them in two years.
Here is the throughline the report exposes. Accurate search, real inventory visibility, honest pricing, and confident recommendations are not five separate initiatives. They are five outputs of the same input: the quality and structure of your product data. When search returns the wrong item, when availability is unclear, when a specification is missing, the shopper experiences a broken basic. Underneath it is a catalog that was never built to answer the question being asked.
AI shopping agents make this visible in a way a website never did. An agent reads product data directly and answers from what it finds. If the data is not there, the agent does not infer it. It works with whatever is available, which means thin or inconsistent product data turns into a wrong answer or no answer at the exact moment a shopper is deciding. The industry has underinvested in the catalog layer for years because a human browsing a page could paper over the gaps. Agentic commerce removes that cushion. The basics consumers say are missing and the product data agents cannot read are the same problem seen from two directions.
Deloitte identified a group it calls “commerce standouts”: brands in the top quartile of execution across search, selection, fulfillment, service, and loyalty. They did not win with a single feature. They executed consistently on the fundamentals, added data-driven personalization, and put AI to practical use.
The payoff shows up where it counts. Standouts are 2.9 times as likely as low-maturity brands to say customers would rate their experience excellent, twice as likely to differentiate on experience, and they outperform on profitability. And when a brand does deliver consistently great experiences, 90% of consumers respond by buying more, recommending the brand, or engaging with it more.
The report frames personalization and AI as the next moves, and names shopping assistants and online search among the top personalization priorities for 2026. Both of those run on product data. A shopping assistant can only be as good as the catalog it reads, and personalized search cannot recommend what it cannot interpret. Getting the fundamentals right is not separate from the AI agenda. It is the foundation the AI agenda depends on.
The uncomfortable part of the Deloitte study is not that brands are failing. It is that many do not know they are, because their own dashboards report the layers that are working while the gap widens underneath. The first move is not a new campaign. It is an honest look at how your product data actually performs for the agents and shoppers reading it, compared with how good you assume it is.
That honest look is what Agentic Commerce Optimization (ACO) is built to give you: making product data complete, structured, and interpretable so the basics consumers care about hold up wherever they shop. This is not a case for tearing out SEO. SEO still governs how you show up in search, and it should keep running. ACO is the layer it never covered, the one that decides whether an agent can include and recommend your products. The two run in parallel and compound. Investing in SEO was never the error. Treating it as the whole scoreboard is. The brands that close the perception gap will be the ones that stopped trusting a single scorecard and started measuring the one their customers, and the agents shopping for them, actually use.
See how AI shopping agents currently read your catalog with ReFiBuy's Agentic Readiness Report, a free diagnostic on where your product data is complete, where it is not, and where the gap between perception and reality is hiding.