You do not need a special trick to get your business recommended by an AI assistant. You need the plain facts about you (what you do, where, when, and what customers say) to be true, current and the same everywhere an assistant or a customer might look. Google, the one maker that publishes guidance, says no special optimization is required.
That is the short version, and it runs against a lot of what is being sold to owners this autumn. The long version is below: who is asking AI to find businesses, what Google actually says, three ways owners respond, a one-rule test, and a five-question check you can run before lunch.
Are customers really asking AI to find local businesses?
Yes, in numbers large enough to take seriously and small enough to keep your head. Yext, a company that sells local listings software, surveyed 3,848 adults in the US, UK, France, Germany and Italy in spring 2026. Its report found that 44% had used an AI tool for local search in the past month. The figure was 47% in the US (1,120 respondents) and 38% in the UK (600). And 29% had tried a new local business because an AI tool recommended it in the previous six months.
Read that with the right amount of salt. The survey was online and self-reported, the report gives no exact fieldwork dates, and Yext itself calls the findings directional.
The more useful finding is what customers do next. According to the same report, 95% take at least one extra step after an AI answer. 53% search Google or Bing, 49% visit the business's website, and 28% check reviews. My reading, and it is only a reading: the assistant introduces you, and then the customer checks you. Your website and your listings are the second half of every AI recommendation, so a weak page wastes a good introduction.
What does Google actually tell owners to do?
Google says there is nothing extra to do. Its page on AI features and your website, last updated 10 December 2025, states: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." A page only needs to be indexed and eligible to appear in Search with a snippet.
A longer guide, Optimizing your website for generative AI features on Google Search (last updated 10 July 2026), goes further. It tells site owners to "focus on developing unique, expert-led content that provides value beyond common knowledge." It says you "don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search." It warns that seeking inauthentic "mentions" across the web "isn't as helpful as it might seem." And it names Google Business Profiles as a way for local offerings to be visible in AI responses as well as in ordinary results.
Now the limit. That guidance covers Google's own AI features. Yext's survey puts ChatGPT first among AI tools its respondents use, followed by Google Gemini and Microsoft Copilot. I did not find an equivalent published statement from the makers of those tools this week. So for them, "do the basics" is my inference, not a finding. It is a sensible inference, because every assistant that recommends a business has to get its facts from somewhere, and the somewhere is mostly web pages and listings.
Which response to AI search holds up?
Owners tend to react in one of three ways. I checked each against the sources above on 7 November 2026.
| Approach | What the sources say | Fits when | Breaks when |
|---|---|---|---|
| 1. Fix the plain facts, then publish what only you know | Google: no special optimization; build "non-commodity" content; keep business details accurate. Customers verify on your website and listings (Yext). | Nearly every business with customers who search before they buy | Your facts are fine but your site says nothing a competitor could not say |
| 2. Add AI-specific files and formats (an llms.txt file, special markup, pages chopped into chunks) | Google says none of it is needed for Google Search. I found no claim from other assistants' makers either way. | You already have a developer and an hour to spare, after approach 1 | It replaces approach 1, or someone sells it as a guarantee |
| 3. Buy mentions or reviews | Google says inauthentic mentions are not as helpful as they seem. US and UK rules ban fake reviews (next section). | Never | Always |
The verdict is plain. Approach 1 is the work. Approach 2 is optional garnish, and harmless if it costs you little. Approach 3 is a risk with no evidence behind it.
Are fake reviews and bought mentions worth the risk?
No, and the risk is legal as well as practical. In the US, the Federal Trade Commission's rule on consumer reviews and testimonials has applied since 21 October 2024. Its staff guidance (last updated 1 May 2025) lists writing, selling or buying fake reviews, paying for reviews of a particular sentiment, undisclosed reviews by insiders, and suppressing honest reviews as violations. The rule lets courts impose civil penalties for knowing violations, and the guidance describes no small-business exemption.
In the UK, the Digital Markets, Competition and Consumers Act changes came into force on 6 April 2025. The Competition and Markets Authority's guidance for businesses (last updated 3 December 2025) bans commissioning fake reviews and says a trader who publishes reviews must take appropriate steps to make sure they are genuine. Fines can reach 10% of worldwide turnover.
Asking real customers for honest reviews is fine, and the FTC guidance says incentives are not banned as long as nothing requires the review to say something particular. This is general information, not legal advice.
What is the Plain Facts Rule?
The Plain Facts Rule is one rule for being found by AI assistants: before you try anything clever, make sure the plain facts about your business, what you do, where, when, for whom and what customers say, are true, current and identical everywhere a customer or a machine might look. Assistants can only recommend what they can confirm.
The facts live in four places: your Google Business Profile, your own website, the directories and listings that copied you years ago, and the social and review pages customers read. Google's profile guidelines ask for your real-world name, an accurate address or service area, your regular customer-facing hours, and "as few categories as possible" that describe what your business is.
Take a hypothetical 12-person plumbing and heating firm. Its Google profile says closed on Sundays. Its website promises 24-hour emergency callouts. An old directory lists a phone number from before the office moved. Ask an assistant who to call on a Sunday and it has three contradictory answers, so it may name a competitor with one consistent answer. Nothing about that firm needed cleverness. It needed an hour with all three pages open side by side.
This is the same idea as the Single Source Test in is your business data ready for AI, pointed outward at the places customers look.
What I Tell Business Owners
Across the owners I speak to, the first question about AI search is what to buy. I have trained more than 150,000 people across 30+ countries, and the people who ask it are rarely short of tools. They are short of an hour to look at what is already out there about their own business.
So my answer is to spend that hour before spending anything. My governing principle applies here: outsource the doing, not the thinking. An assistant can draft the hours, the service descriptions and the answers to customer questions. A person decides what is true, because a confident wrong answer repeated on four sites is worse than one gap.
Then add the thing Google keeps asking for: one page that says something only you can say. Not "we offer quality service." A real answer to the question customers ask you most, with the specifics a competitor would have to invent.
How do you check what assistants say about your business?
Use the Ask-Back Check. It takes about 30 minutes and needs no tool you do not already have.
- Pick two assistants your customers might use, such as ChatGPT and Google Gemini.
- Ask each: "Who is a good [your trade] in [your town]?"
- Ask: "Is [your business] open on [a day you are unusual about]?"
- Ask: "What does [your business] do, and who is it for?"
- Ask: "Does [your business] offer [your most specific service]?"
- Ask: "What do customers say about [your business]?"
Write down three things for each answer: were you named, were the facts right, and did it point to a page you control. Then repeat the questions on a different day. Expect answers to vary, because these tools do not return the same words every time, so look for patterns, not single answers. A wrong fact that shows up twice is a job. A missing mention on one day is noise.
Fix wrong facts at the source, not in the assistant. If the hours are wrong in the assistant, they are wrong somewhere it reads.
What should you do this week?
- Monday, 30 minutes. Write the plain facts on one page: name, address or service area, phone, hours, services, and three things customers say. Pick one person to own it.
- Tuesday, one hour. Open your Google profile, your website and your two biggest directory listings. Fix every mismatch. Use the profile guidelines above for the name, hours and categories.
- Wednesday, 30 minutes. Run the Ask-Back Check and write down the results.
- Thursday, one hour. Add one page or section that answers your most common customer question with details only you know.
- Friday. Put a review date four weeks out in the calendar and repeat the Ask-Back Check then. Measuring whether AI is working covers how to judge a change like this without fooling yourself.
What if an agency promises to get you into ChatGPT?
Treat it like any supplier that says it uses AI on your behalf. Ask which pages and listings they will change, who owns the accounts afterward, how they will show a result, and whether any of it involves buying reviews or mentions. If the answer to the last one is yes, stop. My supplier questions work for this conversation too, and the pitch itself is worth a look through the lens of AI scams aimed at small businesses.
Nobody outside the assistants controls what they say. Anyone who promises a place in the answer is selling a hope.
Where to go next
If you run a shop, restaurant or trade business, the same facts feed your day-to-day AI use, as in AI for retail and hospitality and AI for trades and construction. For the customer-facing side, see AI for customer service and AI for sales and marketing. If browser agents that act for customers interest you, AI agents for small businesses is the place to start.
If you want to know where your business stands before you spend another month experimenting, take the Business AI Readiness Scorecard. If your team needs to learn to check, correct and use these tools well, that is what my AI training for teams is built for. And if you are planning an event for owners in your area, I speak on this at business keynotes.
Sources and further reading
- Yext, The Tipping Point is Already Here: What Yext's 2026 Consumer Search Survey Reveals About AI and Local Discovery, spring 2026 (exact publication and fieldwork dates not stated; 3,848 adults in five countries; online, self-reported).
- Google Search Central, AI features and your website, last updated 10 December 2025.
- Google Search Central, Optimizing your website for generative AI features on Google Search, last updated 10 July 2026.
- Google Business Profile Help, Guidelines for representing your business on Google (undated; read 7 November 2026).
- US Federal Trade Commission, The Consumer Reviews and Testimonials Rule: Questions and Answers, published 8 November 2024, last updated 1 May 2025.
- UK Competition and Markets Authority, What businesses need to know about unfair commercial practices, GOV.UK, first published 4 April 2025, last updated 3 December 2025.
Dan Fitzpatrick is a keynote speaker, Forbes contributor and five-times bestselling author who has trained more than 150,000 educators across 30+ countries. About Dan.
