The Interactive Advertising Bureau found that 65% of surveyed recent AI-shopping users said they felt more confident in recommendations informed by credible creator reviews. For businesses, the result gives evidence from real product use a place alongside specifications and price in their marketing plans.
The digital advertising trade group published the research on September 9, 2026. The survey, conducted through Attest in May, covered 2,200 people in the United States, United Kingdom, Australia, Mexico and India. All had used an AI tool for shopping-related research in the previous three months.
The findings concern people already using AI for shopping research. The 65% figure describes reported confidence, not an increase in purchases or the probability that a product will appear in an AI answer.
Businesses can respond by giving customers clearer evidence to evaluate. That includes the exact product being sold, the conditions behind performance claims, the experience of people who have used it, and the commercial relationships behind endorsements.
Specifications should answer buying questions
A product page needs to explain whether an item fits the buyer’s intended use. Dimensions, materials, compatibility, included accessories and operating requirements give customers something concrete to compare.
For a laptop, that might mean identifying the exact processor, memory, ports and model year. For a bag, it could mean internal dimensions, empty weight and whether a particular device fits inside. These details need to apply to the variant being offered, rather than another version in the same range.
Performance claims need their conditions attached. A battery-life figure should identify the test or usage conditions behind it. A claim about load capacity should state the relevant limits. Describe a subscription charge, required accessory or compatibility restriction where it affects the purchase.
Such information also provides material that AI shopping tools may use. OpenAI’s shopping documentation says ChatGPT considers product information from first-party and third-party providers, alongside other content, when selecting results relevant to a user’s request. It may also summarize reviews from public websites, although OpenAI does not verify those reviews and ratings.
A business cannot control how every tool combines those sources. It can check that its own product pages, manuals and retailer listings describe the same model consistently, and correct conflicting specifications.
Real use gives reviews substance
Creator content is more informative when the audience can see what was tested and how. A review should identify the product version, describe the tasks performed and explain how long the reviewer used it.
A short demonstration can establish that a feature works in the conditions shown. It provides less evidence about durability, long-term running costs or performance in other settings. Marketing teams should keep those limits visible when reusing a review or quoting its conclusions.
Photographs of the actual product, recorded demonstrations and clearly described comparisons let buyers inspect the evidence behind a recommendation. If a test measures performance, publish the method and relevant conditions alongside the result.
Acknowledge tradeoffs as part of the description. A lightweight product may offer less capacity; a compact device may have fewer connections. Explain who the product suits and where another option may be a better fit. Avoid turning one reviewer’s experience into a claim about every customer.
Payment and free samples also belong in the account. The US Federal Trade Commission’s guidance for social media endorsements says brand relationships should be disclosed clearly, including when someone receives free or discounted products. Disclosures should accompany the endorsement, rather than appear only on a profile page.
Businesses commissioning reviews should agree on clear disclosures and allow an honest account of the experience. If they reproduce an endorsement on a product page, retain the context that lets customers assess it, including the commercial relationship and any qualifications.
Keep product information current and accessible
Accurate information requires coordination between marketing, product teams and the people maintaining retail listings. Assign responsibility for updating specifications, prices, availability and policy changes across those channels.
Google’s guidance for its AI search features recommends making important information available as text and keeping structured data consistent with the visible page. Structured data is information labeled in a format that software can read, such as a product’s name, price or availability.
Google says its AI Overviews and AI Mode require no special additional optimization beyond its established search practices. Meeting its requirements still does not guarantee that a page will be indexed or shown.
Businesses should also check what a customer sees after following a recommendation. OpenAI warns that changes to prices or shipping terms may take time to appear in ChatGPT’s shopping information. Our earlier coverage examined how inaccurate AI shopping advice can weaken customer trust.
Review the questions reaching customer support and the reasons buyers give for returns. Repeated uncertainty about fit, compatibility or included equipment identifies information worth adding to a product page. Track whether those inquiries and related returns decline after the correction, while accounting for changes in the products sold and customer mix.
