ACTION 11 OF 16
How to get reviews where AI reads them: which sites, how many, how to ask
SHORT ANSWER
Find the 2–3 review sites AI trusts in your niche and fill out your profile on each one completely. Ask every client for a review after each finished project. Your goal: at least as many reviews and at least as high a rating as the competitors AI recommends next to you.
- Time: 15 min a week
- Owner: Account manager
- Impact: ≈ +3–7% Coverage ¹
Kyrylo Poltavets, Founder of Dabudai · LinkedIn
Published · Updated
How do reviews affect AI recommendations?
Before AI search recommends a company, it checks what others say about it: review sites, directories, and the company’s Google Business Profile. In local niches, Google Business Profiles account for 42–59% of all links in AI answers. [study]
Models compare ratings and review counts across options, and those outweigh even brand recognition. [arXiv, 2026]
How to get reviews AI can see, step by step
- Find out which review sites AI relies on in your niche. Use the prompt below, or check the Sources section on the platform for free.
- Pick 2–3 of them and fill out your profile completely: description, services, prices, photos. Keep it identical to your website and other listings (see Make your company description consistent everywhere).
- Write down the rating and review count of the 2–3 competitors AI recommends next to you. That’s your target.
- After every finished project, send the client a direct link to leave a review. Ask everyone, not just happy clients.
- Once a month, compare your rating and review count with your competitors’.
Pair this with: Add your rating to comparison pages · Get into third-party “best of” lists
How many reviews do you need for AI to recommend you?
There’s no universal threshold. Your benchmark is the competitors AI recommends next to you: you need at least as many reviews and at least as high a rating.
A profile with no reviews does almost nothing. One study found that every service ChatGPT recommends in software categories has reviews on Capterra and G2. [study] For local businesses, Google explicitly lists review count and rating among its ranking factors in Maps, [Google] and Google’s AI answers rely on the Google Business Profile.
How far ahead of a competitor do you need to be? In one experiment, AI chose between a well-known brand and a little-known one with identical specs. The little-known brand started winning half the time as soon as it had any one of these three: [arXiv, 2026]
1.6x
more reviews
Source: arXiv, 2026 ↗
+0.075 stars
higher rating
Source: arXiv, 2026 ↗
7.3%
lower price
Source: arXiv, 2026 ↗
Together, rating, reviews, and price account for about 82% of the model’s choice. The brand name accounts for 1%. [arXiv, 2026]
The experiment used products: cosmetics, cables, and batteries. There’s no separate data for services yet.
Which review sites does AI need to see you on?
It depends on your niche and country. For B2B SaaS, a profile on G2 and Capterra is close to mandatory. [study] For a dentist in Austin, G2 does nothing. What counts there is the Google Business Profile and medical directories with reviews, like Healthgrades and Zocdoc.
57%
of links in AI answers about dentists point to a Google Business Profile
Source: study ↗
42–59%
Google Business Profile’s share in other local niches, from lawyers to vet clinics
Source: study ↗
In local niches, the Google Business Profile is the top source in every industry. Next come niche sites: Tripadvisor and OpenTable for restaurants, legal directories for lawyers, medical directories for doctors. The data was collected in the US, the UK, and Australia. [study]
| Niche | Where AI looks most often |
|---|---|
| SaaS and B2B software | G2, Capterra, GetApp, Trustpilot |
| Agencies and B2B services | Clutch, GoodFirms, Google Business Profile |
| Local businesses | Google Business Profile, Apple Business Connect (Apple Maps), Bing Places, Facebook |
| Healthcare | Google Business Profile plus doctor and clinic directories, such as Healthgrades and Zocdoc in the US |
| Hotels and travel | Tripadvisor, Booking.com, Google |
| Ecommerce | Trustpilot, marketplace reviews such as Amazon and Best Buy, Google |
| Legal services | Google Business Profile; in the US, Avvo and Martindale; for firms, Legal 500 and Chambers |
Use this as a starting point. To get the exact list for your niche and city, run the prompt below.
How do you find out which review sites matter in your niche?
Fill in your details and run the prompt 2–3 times in ChatGPT with search and in Google AI Mode. Keep the sites that show up every time.
You are a business reputation analyst. My business: [name], [what we do], [city, country (e.g., Austin, TX, US or Manchester, UK)], website [domain].
Task: figure out which review sites, directories, and aggregators you rely on when you recommend companies in my niche, and which of them you trust most.
1. Search the web for answers to 5 typical customer queries in my niche, for example: “best [service] in [city]” (like “best dentist in Austin”), “[service] [city] reviews,” and “who to hire for [task].”
2. For each query, list the review sites, directories, and aggregators you link to or take ratings from (for example, Google Business Profile, Yelp, Trustpilot, or niche directories).
3. Build a table: site | which queries it shows up for | whether [name] is listed there | our rating and review count | rating and review count for the 2–3 competitors you recommend.
4. Mark the sites people trust most in this niche and country, and explain why.
5. Name the 3 sites I should start with, in priority order.
Only cite real pages with their URLs. If there’s no data, say so. Don’t make anything up.How do you ask a client for a review?
Send it right after the project wraps up, with a direct link to the review form on the site you’re targeting. Ask for an honest rating, whatever it is.
Hi Emily,
Thanks for choosing us for [project].
If you have 2 minutes, would you share how it went on [site]? Here’s the link: [link].
Any rating helps, including a critical one.
Thanks,
[Name]What’s not allowed when you collect reviews?
The FTC’s 2024 rule (US) explicitly bans:
- buying reviews or writing them on behalf of clients or made-up people;
- offering a discount or gift only for positive reviews (an incentive for a review with any rating is fine);
- reviews from employees and relatives that don’t disclose the connection;
- threatening or pressuring people to remove a negative review, or showing only positive reviews as if they were all of them.
The fine is more than $50,000 per violation. [FTC] Review sites have their own rules too. Google, for example, bans asking only happy customers for reviews. So ask everyone after a finished project. [Google]
How do you track where AI sees you and your competitors?
If you use Dabudai, open Sources and filter for review domains. You’ll see which ones AI cites for your queries, and where competitors are mentioned but you aren’t yet.
app.dabud.ai · Sources

No account yet? Your first brand data shows up within 20 minutes of signing up. The first 7 days are free.
Common questions about reviews and AI
Can I buy reviews or ask friends to write them?
No. In the US, the FTC fines this at more than $50,000 per violation, and review sites remove those reviews and profiles.
Can I ask only happy clients for reviews?
No. Ask everyone after a finished project. Most review sites ban selective requests.
Do I need a G2 profile if I run a local business?
No. In local niches, AI usually takes reviews from Google Business Profiles and niche directories in your country. G2 and Capterra matter for software.
What Dabudai customers say
★★★★★
A clear direction, even before the recommendations
“The dashboards are clear and intuitive. They make it easy to see what’s driving our AI visibility.”
Andrew Shestopalov
· Head of Marketing, Hotline
GO DEEPER
Learn more about reviews in Lesson 7.2 of the Learning Center inside the Dabudai platform, with examples by niche and a check on your own data.
The Learning Center is open during the free 7-day trial. No credit card needed.
Sources
[FTC] FTC: Trade Regulation Rule on the Use of Consumer Reviews and Testimonials (2024) ↗
[study] Quoleady: G2 and Capterra reviews and ChatGPT recommendations, 2026 ↗
[arXiv, 2026] Study of well-known brand bias in LLM choices, 2026 ↗
[Google] Google: Maps user-generated content policy ↗
[study] BrightLocal: Which directories and review sites AI cites for local businesses, 2026 ↗
[Google] Google: How local ranking works ↗
¹ Estimated impact comes from Dabudai’s analysis of 350,000 AI answers across 2,300 websites, October 2026. Your results depend on your niche, competitors, and starting point.