Does Review or AggregateRating Schema Help AI Trust a Business?

What does review markup do in Google?

Google's review snippet documentation covers two types. Review is one person's evaluation of an item. AggregateRating is the average of many. When a page qualifies, Google may show star ratings in its results. The rules are specific:

• Eligible items include books, courses, events, movies, products, recipes and software apps.
• Local businesses and organisations are eligible only on sites that collect reviews about other businesses. Reviews of a business shown on that business's own site are self-serving and not eligible. That applies whether the reviews sit in your markup or come through an embedded widget, such as your Google or Facebook reviews.
• Reviews must be visible on the marked-up page, and an aggregate rating must be shown to users.
• Ratings must come directly from users. Do not aggregate ratings from other websites.
• Reviews must be about a specific item, not a category or a list of items.

Markup that breaks these rules does not earn stars and risks a manual action for structured data.

Does adding AggregateRating markup help AI include social proof in answers?

There is no evidence that it does for AI answers generally. Look at the routes one at a time:

• Assistants fetching your page. SearchVIU tested ChatGPT, Claude, Perplexity, Gemini and Google's AI Mode in October 2025. When they fetched pages directly, they extracted visible text and ignored JSON-LD. An aggregateRating block alone would not be seen. A visible line such as "Rated 4.7 from 1,240 reviews" would.
• Search indexes behind AI answers. Google and Bing parse structured data for their own features, and Microsoft has said schema helps its language models understand content. Neither has said review markup affects which businesses AI answers recommend.
• Measured results. Ahrefs compared 1,885 pages that added structured data between August 2025 and March 2026 with matched pages that did not. The changes in ChatGPT and AI Mode citations were indistinguishable from zero. The study pooled schema types, but it is the best controlled evidence available.

AI answers that mention ratings usually quote them from review platforms, such as "4.6 stars on Trustpilot", or from articles that report them. For recommendation questions, assistants often lean on independent validation as well as what a business says about itself, as covered in why AI engines trust independent third-party validation more than brand-owned content.

How can businesses make reviews machine-readable?

1. Ask every customer for an honest review in their own words on the platforms your buyers and AI answers use, such as Google Business Profile, industry review sites and marketplaces. These pages are what answers cite. Yelp is the exception: it does not allow businesses to ask for reviews, so there keep your listing accurate and reply to reviews instead.
2. Check those platforms are readable by AI crawlers. Some block them. Yelp's robots.txt, for example, blocks the main AI search crawlers from most of the site. Trustpilot's review pages are open. A crawler block does not settle the question on its own: Yelp blocks these crawlers but licenses its reviews to OpenAI and Perplexity, and is one of the sources ChatGPT cites most for local businesses. Check which review sources actually appear when you run your buyers' questions, and keep those healthy.
3. Show reviews on your own pages as HTML text: the reviewer's first name or role, the date, the rating and the review itself.
4. State your ratings in plain sentences: "Rated 4.7 out of 5 from 1,240 reviews on [platform], as shown on our profile." Link to the source so readers and systems can verify it.
5. Add markup only where eligible, such as on product and software pages with reviews collected from your own users, and only for what the page shows.
6. Keep review counts and ratings current. An old figure repeated on many pages becomes a stale fact in answers.

Should you paste customer reviews as text instead of using JavaScript widgets for AI crawlers?

Showing reviews as text is better for AI crawlers, but "pasting" needs care. A 2024 study by Vercel and MERJ found none of the major AI crawlers rendered JavaScript. That included the crawlers used by OpenAI, Anthropic and Perplexity. A review widget loaded by script, or inside an iframe, is effectively blank to them. Server-rendered review text is not.

Before copying reviews from another platform onto your site:

• Check the platform's terms. Some allow displaying reviews with attribution, some only through their own widgets, and some not at all.
• Keep the review exactly as written, with the reviewer's name as shown and the date.
• Do not select only the best ones while implying they are typical. The FTC's guidance on its reviews rule says using non-representative reviews in marketing could be deceptive.
• Do not mark copied reviews up as your own ratings. Google's guidelines say not to aggregate reviews from other websites.

A common approach is to keep the platform's widget for the full, live list. Add a short server-rendered summary above it: the overall rating, the count, the source, and a few recent reviews quoted in full.

Should I use Review schema on a first-party software alternatives page?

Usually not. A page on your own site listing alternatives to a competitor, or comparing several products, runs into three problems:

• It is a list. Google says review information should be about a specific item, not a list of items. So the page is not eligible for review snippets.
• The reviewer is not independent. Your ratings of competitors on your own page are an interested party's opinion. Marking them up as reviews presents them as more neutral than they are.
• Mixed-up ratings mislead. Borrowing competitors' ratings from review sites and marking them up breaks the rule against aggregating ratings from other websites.

What does work is a clear comparison table with facts that can be checked, such as prices, limits, integrations and platforms. Cite the sources, say plainly that it is your comparison, and state where each product fits best, including where yours does not. That is more useful to readers and to AI answers than star ratings you awarded yourself. If you review a single competitor's product on its own page, as an editor would, Review markup can be valid. Keep it honest and visible.

What do AI answers use as trust signals instead?

• Review volume, rating and recency on independent platforms.
• What reviews say, summarised. BrightLocal's 2026 research found half of consumers trust AI to summarise reviews, so recurring themes in reviews become the answer's description of you.
• Responses to reviews, which show that problems get resolved.
• Consistency between what your site says and what reviewers report.
• Mentions in independent articles and comparisons.

Clear, accurate business details still matter, as covered in whether LocalBusiness or Service schema helps AI understand business details.

What rules apply to reviews on your site?

• The FTC's rule on consumer reviews, in force since October 2024, bans fake reviews, including AI-generated ones. It bans incentives conditional on positive sentiment. It bans suppressing reviews through threats or intimidation, and it carries civil penalties.
• The UK's ban on fake reviews under the Digital Markets, Competition and Consumers Act took effect in April 2025.
• Google's guidelines exclude self-serving local business reviews from rich results. Markup that does not match the page risks a manual action.

Markup that says more than the page shows is also a trust problem in its own right, as covered in whether schema that conflicts with page content can reduce AI trust or visibility.

What might an illustrative case look like?

Consider an illustrative accounting practice with 180 Google reviews averaging 4.8 stars. It displayed them only through a script-loaded carousel on its homepage, marked up with AggregateRating. AI answers about accountants in its town cited a directory where it had three old reviews. The practice removed the ineligible markup and added a server-rendered line stating its Google rating and review count, with a link. It quoted six recent reviews in full, with dates, and began asking clients to review it on the directory too. Two months later, answers began describing it as "highly rated on Google". One quoted a recurring theme from its reviews, fast responses during tax season.

What do people ask about review schema and AI?

Will star ratings in Google help AI answers mention us?

Stars can raise clicks from search results. There is no evidence they change what AI answers say about you.

Can I mark up testimonials on my homepage?

For a local business or organisation, reviews of yourself on your own site are self-serving and not eligible. Show them as text instead.

Are AI-written reviews allowed?

Not as reviews. The FTC's rule treats reviews that misrepresent who wrote them, or their experience, as fake.

Which review platform matters most?

The one answers for your category cite most. Check by running your buyers' questions and noting the review sources that appear.