Summit found that 41% of leading UK retailers lack the structured data AI shopping agents use to interpret websites. The agency based its findings on an audit of 200 retailers across four consumer categories.
The study covered beauty, fashion, home and garden, and pet care, assessing how well retailers' sites are prepared for AI-driven shopping. It focused on whether machines can read and act on information across homepages, category pages, product pages, and blog or content pages.
In addition to 41% missing key structured data, 20% lacked descriptive image alt text, and 8% had gaps in product and category content. Structured data was the most common issue in every category reviewed.
Beauty retailers recorded 20 structured data issues, followed by eight heading issues and seven alt text issues. In fashion, the audit flagged 26 structured data issues, 10 content issues, and seven alt text issues. Home and garden retailers showed 19 structured data issues, 15 alt text issues, and six content issues, while pet care retailers recorded 17 structured data issues, 11 alt text issues, and seven heading issues.
The scorecard evaluates nine technical areas that affect how easily a site can be interpreted by machines. These include crawler access, structured data, content depth, what Summit described as LLM trust signals, heading hierarchy, metadata, internal linking, image alt text, and accessibility labelling.
Summit paired the technical review with AI visibility and sentiment data from Searchable, a conversational search platform. The assessment used 50 prompts per retailer on a single AI platform and captured a single point in time.
Adam Preston, Head of SEO at Summit, said: "The gap between what looks good to a shopper and what parses cleanly for a machine is massive. AI agents don't browse websites; they process data. If an agent can't deterministically read your product attributes, it simply moves on to a competitor."
The audit treats crawler access as a gatekeeper measure, meaning a page that cannot be accessed by an AI crawler cannot be assessed on the other dimensions. The readiness scores reflect a sample of key pages rather than a full-site crawl.
Why data matters
Structured data helps machine systems identify the most important elements on a page, such as pricing, availability, product features, or article content. In practical terms, gaps in those markers may limit whether products appear in AI-generated answers, comparisons, or shopping recommendations.
Preston said: "Structured data acts as flags to search engines and LLMs as to what the most important bits of information are about that page. These could be product features, price or blog content. If this isn't complete, it limits the range of searches and prompts you'll be eligible to show for, as well as restricting the number of search features that your pages may appear in."
The scorecard marks the return of a benchmark Summit first introduced a decade ago, when retailers were adapting to mobile commerce. The latest version shifts that framework towards AI-led product discovery and comparison.
Martin Corcoran, CEO at Summit, said: "We brought the Scorecard back after ten years because retail is undergoing its biggest structural shift since the smartphone. The way people discover and choose products is changing again, and AI is becoming an increasingly important part of that journey. Today, many household brands can still appear in AI-generated answers because they have strong brand recognition, established content and years of online authority behind them. But as AI moves beyond answering questions and towards helping people compare products and complete purchases, that alone may not be enough. The good news is that this isn't an abstract problem - technical readiness can be measured, understood and improved. That's why we felt it was important to bring the Scorecard back now: to give retailers a clear view of where they stand and what they can do next."