Blog

Do online reviews affect AI recommendations?

By Dawn, Founder of CitedAsia · Published 5 November 2026

Customer paying a shopkeeper at a small shop counter
Short answer: Yes, very likely, though no AI company publishes an exact formula. Google states that more reviews and positive ratings can help local ranking, and its Maps AI draws on community ratings and reviews; ChatGPT and Perplexity build answers from web pages, which often include review sites. The practical takeaway: earn a steady stream of genuine, detailed reviews on the platforms that matter for your category, ask within Google’s rules, and reply to every review.

Business owners in Singapore and Malaysia often ask whether reviews still matter now that customers ask AI tools for recommendations. The honest answer is that reviews probably matter more, not less, because an AI answer often has to justify why it named you. “Highly rated for punctual service” has to come from somewhere. This article separates what is documented from what is likely, then covers how to collect and handle reviews safely.

What do we actually know about reviews and AI recommendations?

Google has documented that reviews affect local ranking and feed its Maps AI; for ChatGPT and Perplexity, the link is indirect but logical.

Here is what the sources say, as of October 2026:

ClaimStatusBasis
Reviews and ratings help Google local rankingDocumentedGoogle Business Profile Help
Google Maps AI uses ratings and reviewsDocumentedGoogle blog, Feb 2024
ChatGPT and Perplexity cite pages that include reviewsObservableCheck the Sources list on any local question
Detailed reviews give AI more to quote than “Great!”LikelyAI answers summarise text; short reviews carry little information
A specific review count guarantees AI mentionsUnknown / falseNo AI company publishes thresholds

Which review platforms matter in Singapore and Malaysia?

Google comes first for almost every local business; after that, it depends on your industry and on which sites AI tools cite for your category.

Google reviews show up in Search, Maps and Google’s AI features, so start there. Then look at the platforms specific to your field. A Singapore renovation firm may find Qanvast, which lists interior design firms with ratings and reviews, in AI sources. A Malaysian home service business may see Recommend.my, whose pro profiles include customer reviews. Restaurants, clinics and tuition centres each have their own.

The way to know for sure is to check. Ask ChatGPT (with search on) and Perplexity the questions your customers ask, open the sources, and note which review sites keep appearing. Focus your effort on two or three of those rather than spreading thin across ten.

The review cycle: serve, ask, customer writes, owner replies, AI can quote, with three things to avoid A review cycle that stays within Google’s rules 1. Serve well 2. Ask everyonelink or QR code 3. Customer writesin their own words 4. You replyshort and specific 5. Details to quoteservice, area, outcome Avoid ✕ Incentives Discounts, gifts or vouchersfor leaving a review ✕ Asking only happy customers Google calls this selectivelysoliciting positive reviews ✕ Staff quotas or staff reviews Quotas and conflict-of-interestreviews break the policy Source: Google contribution policy
Every step in the loop adds detail an AI tool can use; the three items on the right are against Google’s policy.

How can I ask for Google reviews without breaking the rules?

Ask every customer the same way, using Google’s review link or QR code, and never offer anything in return.

Google’s own advice is to ask customers to visit a Google link or scan a QR code. Its fake engagement and rating manipulation policies prohibit:

  • Offering payment, discounts or free goods or services in exchange for a review, or for changing or removing a negative one
  • Discouraging negative reviews, or selectively soliciting positive reviews from customers
  • Requiring staff to collect a set number of reviews
  • Reviews based on a conflict of interest, which can include current or former employees

That rules out common tactics like “free drink for a 5-star review”, or texting the review link only to customers who said they were happy. A simple, compliant approach: send the review link to every customer after the job, by WhatsApp or email, with one polite line. Google itself notes that a mix of positive and negative feedback often feels more trustworthy.

QR code stand on a cafe table
A review link or QR code, offered to every customer, keeps you within Google’s rules. (Photo: Unsplash)

What makes a review more useful to an AI tool?

Detail. A review that mentions the service, the location and what happened gives an AI something concrete to repeat.

“Great service!” tells an AI almost nothing. “They fixed the leaking pipe in our Punggol flat the same afternoon and explained the cost before starting” tells it the service, the area, the speed and the pricing behaviour. You cannot write reviews for customers, and you should not give them a script. But you can make the request easier to answer well:

  • Send the link soon after the job, while details are fresh.
  • Ask an open question, such as “If you have a moment, we’d appreciate a review of how the job went.”
  • Use your service names consistently on invoices and messages, so customers naturally use the same words.

How should I respond to reviews, including bad ones?

Reply to all reviews briefly, personally and politely; for negative ones, acknowledge the issue and move the detail to a private channel.

Google’s guidance on replies says to keep them short and simple, avoid identical thank-you messages, not use replies to promote deals, never share a reviewer’s private information, and invite the person to continue by phone or email for complex issues. A reply also adds your own facts to the page, for example “We’ve since added Saturday appointments”, which customers and AI tools can both read.

A calm, specific reply to a one-star review often does more for trust than ten generic thank-yous. If a review breaks Google’s content policies, flag it rather than arguing in public.

Try this today: Create your Google review link from your Business Profile, save it as a WhatsApp quick reply, and send it to every customer you served this week, not only the ones you think were happy. Then reply to your three most recent unanswered reviews.

Frequently asked questions

How many reviews do I need before AI tools recommend me?

No AI company publishes a number. Google says more reviews and positive ratings can help local ranking, so aim for a steady flow over time rather than a target count.

Can I offer a discount to customers who leave a review?

Not on Google. Its policy prohibits offering payment, discounts or free goods or services in exchange for reviews.

Should I only ask customers who were happy?

No. Google’s policy prohibits selectively soliciting positive reviews. Ask every customer the same way.

Do reviews on sites other than Google matter?

They can, especially if AI tools cite those sites for your category. Check the sources ChatGPT and Perplexity list for your customers’ questions to see which platforms appear.

Curious whether your reviews are showing up in AI answers? Our free AI visibility check gives you a score in 30 seconds, plus a report within 24 hours (business days) that tests five customer questions on ChatGPT and Perplexity and shows whether your business is named. You can also read what GEO is for the wider picture, or see our pricing.

Is your business in the answer?

Free check of 5 customer questions on ChatGPT and Perplexity. Score in 30 seconds, full report within 24 hours (business days).