Do online reviews influence ChatGPT recommendations?

Online reviews are one of several signals that can influence which businesses ChatGPT recommends, but the evidence we hold supports a narrower claim than "reviews drive recommendations". This FAQ separates what the sources show from what they do not, so you can decide where review effort belongs in your AI visibility plan.

Reviews and AI recommendations

Do online reviews influence ChatGPT recommendations?

Reviews can influence recommendations as one corroboration signal among several, but we can only support that claim at the level of entity consistency, not as a measured ranking factor. Models read your name, address, registration details, people and claims across your site, regulators, directories, review platforms, professional bodies and Companies House, and these must say the same thing everywhere a model can read them. Review platforms are one of the places a model can read about you, so what appears there needs to match the rest of your footprint. We do not have a controlled study that isolates review volume or star rating as a cause of a ChatGPT recommendation, and we will not claim one.

How do reviews fit into third-party corroboration?

Reviews are one form of third-party corroboration, and third-party mentions carry more weight than most businesses assume. In an Ahrefs study of 75,000 brands, branded web mentions showed a 0.664 correlation with AI visibility, against 0.218 for backlinks. That finding is about mentions in general, not reviews specifically. It tells you that being discussed by independent sources matters more than owning links, and reviews on recognised platforms are one way of being discussed.

Do contradictions in my reviews or listings hurt me?

Yes, contradictions are the most common reason a model stays vague about a provider. If a review platform lists a different address, a different service description or outdated registration details from the ones on your website, the model has conflicting evidence and tends to hedge. Audit every platform that carries your business name and make the facts match before you spend effort generating more reviews.

Local and service businesses

Are reviews more important for local businesses?

Reviews matter most in local search, and our own past work gives one concrete example. We took a local estate agent that was not appearing in Google's local listings and got it to number one locally. Calls rose 1,200% from a base of roughly 500 to 600 a month, and ranking there depended on reviews. That result is from Google's local listings in 2019, not from ChatGPT, so read it as evidence for how much reviews drive local visibility on Google rather than as proof of an AI effect.

Does Google local visibility carry over to ChatGPT?

No, and the gap is large. SOCi's 2026 analysis of more than 350,000 locations found that 1.2% were recommended by ChatGPT, against 35.9% appearing in Google's 3-pack. A strong local profile on Google does not translate automatically into being named by ChatGPT. Treat reviews as one input to local visibility and measure AI answers separately.

What to do about reviews

Which review platforms should I focus on?

Start with the platforms your customers already use and a model can already read. The sources we hold name Google, Facebook and Trustpilot as the main review destinations, and recommend asking for a Google review first because Google reviews are the most valuable for search. We have no data ranking these platforms by their influence on ChatGPT specifically, so we do not claim one platform wins for AI answers.

How should I ask customers for reviews?

Follow up every sale and ask for a review, giving people a choice of platform. You can automate the request, but offer Google first, then Facebook or another platform. You can also offer an on-site review option, which brings customers back to your domain and gives you review text you control. Do not pay for reviews with vouchers or cash. Instead, make customers feel part of an experience, for example by offering a piece of useful content in return for feedback.

What about fake reviews?

Fake reviews are a persistent problem, including competitors posting them against you, and there is little you can do to stop them at the moment. On Trustpilot, a premium account gives you an account manager who can get fake reviews taken down, though the process remains difficult. Report fakes promptly and keep your genuine review flow steady so that false entries do not dominate the picture a model reads.

Can I rely on reviews alone to get recommended?

No. Models retrieve and cite content that is structured as self-contained, answer-first passages, with a complete standalone sentence under each heading that can be lifted out and quoted. Reviews do not provide that structure. Reviews support the credibility of what your own pages say, so publish clear, quotable answers on a domain you control and use reviews to corroborate them.

Should I worry about blocked crawlers before I work on reviews?

Check that first, because it is the first binary test in any audit. Cloudflare's default settings block AI crawlers, the default block has been active since July 2025, and roughly 22% of web traffic sits behind Cloudflare. Content behind a default block is invisible to AI engines regardless of how well it is structured. Verify your own settings before you invest in reviews or content.

Measuring the effect

How do I know whether reviews are changing my AI visibility?

Track named citations across a fixed set of prompts, every cycle, on every engine, with screenshot evidence of where you appear and who appears instead. A single check carries a margin of roughly ±9 to 13 points, which is why one screenshot proves nothing. Our own method samples around 1,500 answers per source per cycle across 30 prompts, with Wilson confidence intervals and a published margin of error of roughly ±1 to 2 points.

How do I attribute enquiries to AI engines?

Add a "How did you hear about us?" field to your enquiry form and list the individual AI engines as options. This gives you direct evidence of which engines send you business, which is more reliable than inferring influence from review counts.

What is the practical takeaway?

Make your facts consistent across every platform that mentions you, keep a steady flow of genuine reviews on recognised sites, publish answer-first content on your own domain, confirm AI crawlers can reach it, and measure named citations against a fixed prompt set. Reviews support that work. They do not replace it.