Reviews do not rank you inside an AI assistant, but they heavily shape what it says about you. Assistants summarise the public text written about a business, and for most service businesses in Thailand reviews are the largest body of that text. A profile of recent, specific, written reviews across Google and the platforms your industry uses gives an assistant something concrete to quote. A profile of silent five-star clicks gives it nothing.
What an assistant actually reads
When someone asks ChatGPT or Google AI Overviews for a recommendation, the assistant is not consulting a ranking table. It is assembling an answer from whatever it can read about the businesses it knows in that category and area, then choosing the ones it can describe with some confidence.
Your website supplies the claims. Your reviews supply the corroboration. That distinction matters, because an assistant asked to justify a recommendation will lean on the source that was not written by you.
Evidence an assistant weighs
- Written detail that names a service, a situation or an outcome
- Recency, because an old profile may describe a business that has changed
- Agreement between Google, booking sites and local platforms
- Owner replies, which add context the review itself lacks
- Language, since it answers in the language it can read you in
Why the star rating matters less than you think
Ratings filter. Text decides. By the time an assistant has narrowed a category down to a handful of candidates, most of them sit within a few tenths of a point of each other, and the number stops separating anyone. What separates them is whether the reviews say anything useful.
Consider two clinics with the same score. One has fifty reviews reading "good" and "recommended". The other has fifty reviews mentioning weekend appointments, English-speaking staff, transparent pricing and short waiting times. Ask an assistant which clinic suits a working expat who needs a Saturday appointment, and only one of them can be answered for.
This is the same reason detailed reviews outperform generic ones in ordinary local search. If your listing needs work at that level first, start with Google Maps optimisation and let the AI benefit follow.
Generic vs quotable
- "Great service, recommended." Nothing an assistant can match to a question.
- "Booked a same-day appointment in Thonglor, staff explained the pricing in English before starting." Answers at least four different questions.
The Thailand problem: your reviews speak one language
Thailand is a two-language market, and most review profiles here quietly pick a side. A Bangkok clinic with excellent Thai reviews and nothing in English is hard for an assistant to recommend to a medical tourist asking in English. A hotel with hundreds of English reviews from foreign guests and nothing in Thai has the same problem in reverse when a Thai family asks in Thai.
Assistants answer in the language of the question, and they are more confident when the supporting evidence is in that language. You cannot dictate what language a customer writes in, but you can decide who you ask, and when. If half your revenue comes from a segment that has never left you a review, that is a visibility gap and not just a marketing one.
The same logic applies to your own pages, which is why a properly written Thai version of your site is a GEO investment rather than a courtesy. We cover the wider version of this problem in the guide to SEO for Bangkok businesses.
Where else AI looks for opinions about you
Google reviews are the largest single source for most businesses, but they are rarely the only one. Assistants read broadly, and agreement across independent sources is what turns a claim into something worth repeating.
Which platforms matter depends on your industry rather than your preference. Hotels are read through booking platforms. Restaurants are read through Google, delivery apps and local review sites. Clinics and professional firms are read through Google, industry directories and whatever forums their customers use. The pattern to aim for is simple: the same strengths, described in similar terms, in more than one place.
Sources worth keeping current
- Google Business Profile, the anchor for almost every local business
- Booking platforms such as Agoda or Booking for hospitality
- Local review platforms your Thai customers already use
- Industry directories that assistants treat as reference sources
- Your own site, where testimonials should name the service, not just praise you
How to earn reviews an assistant can use
The goal is not more reviews. It is more reviews that contain information. That changes how you ask.
- Ask at the moment of relief. Right after the problem is solved, not a week later in a bulk email. People describe what happened when they still remember it.
- Ask a question, not for a favour. "Would you mind mentioning which treatment you came in for?" produces detail. "Please leave us 5 stars" produces a click.
- Ask both halves of your market. If your Thai customers review you and your international customers do not, fix the asking process, not the reviews.
- Reply to every review with specifics. Your reply is indexable text you control, and naming the service and the location in a natural sentence adds context the review may lack.
- Keep a steady flow. A few reviews every month is worth more than thirty in one week, which looks engineered to platforms and to readers.
- Answer the criticism properly. A calm, factual reply to a bad review is often the most credible text on your entire profile.
How to check what AI currently says about you
This takes fifteen minutes and does not need a tool. Open an assistant, ask the questions your customers would ask, and read the answers as a stranger would.
- Ask for the best provider of your service in your area, and see whether you appear at all.
- Ask the same question in the other language, and compare the two answers.
- Ask directly what people say about your business by name, and check whether the description is accurate, outdated or generic.
- Ask a question tied to something you actually do well, such as late opening hours or a specialist service, and see whether the assistant credits you with it.
If the answers are vague, the usual cause is not that AI has ignored you. It is that there is nothing specific written about you to work with. That is a fixable problem, and it is the same work that makes you easier to choose for the humans who read the reviews directly. The vertical-specific version of this is worth reading too, for example how Bangkok clinics get named in AI search.
Want to know what AI says about your business?
We will run your business through the assistants your customers use, check what your review profile is telling them, and send you the gaps in plain language.
Get Your Free AI Visibility Check →