Authority 9 min read
Do Reviews Affect AI Recommendations? What the Data Shows, and How to Get More (Legally)
Reviews behave less like a ranking factor and more like a gate. Below a certain rating, or with nothing recent, you rarely make the shortlist an assistant reads out.
Founder, SEO·AEO·GEO
Builds SEO·AEO·GEO and writes the crawler, schema, and citation research published here.
Key takeaways
- In SOCi’s study of 350,000+ locations, the ones ChatGPT recommended averaged 4.3 stars. Ratings act like a threshold.
- Recency matters to people and to the data: 74% of consumers look only at reviews from the last three months (BrightLocal 2026).
- Be on several platforms. SE Ranking found domains listed on several review sites averaged 4.6–6.3 ChatGPT citations, against 1.8 for none.
- Ask every customer, the same way, at the right moment. Google bans incentives and selective asking; the FTC bans fake and bought reviews.
- Reply to reviews, especially bad ones. 89% of consumers expect owners to respond.
Reviews affect AI recommendations mostly as a filter. The businesses AI assistants recommend tend to clear a rating threshold (in SOCi’s data, ChatGPT’s picks averaged 4.3 stars), have recent reviews, and appear on more than one review platform. Beyond that, what reviews say matters more than how many there are, because assistants summarise them. So the work is a steady flow of genuine, recent reviews on Google and the platforms your category relies on, with replies, gathered within the rules that ban incentives, selective asking and fake reviews.
What the evidence says
| Finding | Source | What it suggests |
|---|---|---|
| Locations ChatGPT recommended averaged 4.3 stars; ChatGPT recommended 1.2% of locations against 35.9% shown in Google’s 3-pack | SOCi, 350,000+ chain locations (vendor, correlational) | A threshold, not a race; a top Google ranking doesn’t carry over |
| Domains on several review platforms averaged 4.6–6.3 ChatGPT citations, against 1.8 for domains on none | SE Ranking, 129,000 domains (vendor, correlational) | Spread across platforms, not just Google |
| Ask Maps recommends from 300 million+ places and reviews from 500 million+ contributors | Google, via AP (March 2026) | For Gemini in Maps, reviews are the raw material |
| Yelp licenses reviews and business data to OpenAI | MarTech (July 2026) | Yelp reviews reach ChatGPT through a licence, not crawling |
All of the studies are correlational, and SOCi’s sample is multi-location brands rather than independents. None shows that adding reviews causes recommendations. Together they show that recommended businesses look a particular way: well rated, recently reviewed, present on several platforms.
Which platforms matter
Start with Google, then add the platforms the assistants cite for your category. Ask your customers’ questions in ChatGPT, Perplexity and Copilot, note which review sites appear in the sources, and prioritise those. BrightLocal found consumers use an average of six review sites; for writing reviews, the most used were Google (45%), Facebook (34%), Yelp (24%), Apple Maps (17%), Tripadvisor (16%) and the Better Business Bureau (16%). B2B software has its own (G2, Capterra, Trustpilot); trades and clinics have theirs.
How to get more reviews, within the rules
- Ask everyone, not just the happy ones. Google’s policy prohibits selectively soliciting positive reviews or discouraging negative ones (“review gating”).
- Ask at the moment of satisfaction: when the job is finished, the order arrives or the appointment ends. A short message with a direct link to your review page works better than a generic “find us on Google”.
- Make it one tap. Google Business Profile gives you a short review link; put it in your follow-up email, receipt or text.
- Don’t pay or reward. Google prohibits incentives “such as payment, discounts, free goods and/or services” for reviews.
- Don’t write them yourself or ask staff and family to. Google’s policy covers conflicts of interest, and the FTC’s rule bans reviews that misrepresent who wrote them.
- Spread requests across platforms by rotating which link you send, so one customer isn’t asked for four reviews.
Reply to reviews, especially bad ones
In BrightLocal’s 2026 survey, 89% of consumers expect owners to respond, and 80% are likely to use a business that responds to all its reviews. Replies are also text that sits next to the review, and assistants that summarise a business read both. A specific, calm reply to a complaint (“we’ve changed how we book follow-ups; call me directly”) turns a bad review into evidence that you fix things. How what reviews say shapes the tone of AI answers is the subject of how to get AI to say good things about your business.
Your own site still matters
- Show a few real reviews on your service pages, as text, attributed and dated. Google doesn’t give star rich results to reviews a business publishes about itself, so don’t expect markup to add stars.
- Link to your profiles on the review platforms from your site, and back.
- Keep your business details identical on every platform, so reviews attach to one business, not three versions of it (help AI understand your business).
For the complete local picture (Business Profile, Apple, Bing, Yelp and the city pages assistants cite), read getting a local business recommended by AI.
FAQ
Do Google reviews affect AI Overviews and Gemini?+
For local questions, yes. Google says Business Profiles can help your products and services be visible in AI responses, and Gemini in Google Maps builds recommendations from Maps places and reviews from more than 500 million contributors. Ratings, recency and what reviews say all feed that.
Does ChatGPT read Google reviews?+
ChatGPT reads whatever its search retrieves, which often includes review sites and directories; Google Maps reviews themselves are not simply crawlable pages. Yelp licenses its reviews and business data to OpenAI, so Yelp reviews reach ChatGPT through that agreement. The safest approach is reviews on Google plus the platforms ChatGPT cites for your category.
How many reviews do I need to be recommended by AI?+
There’s no published number. The data points to rating and recency more than volume: SOCi found recommended locations averaged 4.3 stars, and most consumers only read the last three months. A steady flow of recent, genuine reviews beats a large old total.
Can I offer a discount for reviews?+
No, not on Google: its policy prohibits offering incentives, including discounts or free goods, in exchange for reviews. In the US, the FTC’s 2024 rule also bans buying positive or negative reviews. You can ask every customer for an honest review, with no reward attached.
Sources
- In AI-driven discovery, few brands are chosen and most disappear (350,000 locations) — vendor study — SOCi (2026-01-28)
- Local Consumer Review Survey 2026 — BrightLocal (2026-02-11)
- ChatGPT ranking factors (129k domains, correlational) — vendor study — SE Ranking (2025-11-24)
- Maps user-contributed content policy: prohibited and restricted content — Google Maps Help
- Federal Trade Commission announces final rule banning fake reviews and testimonials — US Federal Trade Commission (2024-08-14)
- Optimizing your website for generative AI features on Google Search — Google Search Central (2026-07-10)
- Google unveils major changes to its Maps app (AP) — WNEM (2026-03-12)
- Yelp brings reviews and local leads to ChatGPT — MarTech (2026-07-28)
Run it on your domain
See this lens as a finding, not a lecture.
Free technical GEO audit. Evidence-traced scores and an LLM-ready export. See your score with no login.
Audit your website
