If you have ever asked ChatGPT who to hire for a job, you already know what your customers are doing now. They type a question, get three or four business names back, and often never open a page of search results at all.
So the question worth answering is a simple one: how do you get your business named in those answers?
The short version is that you cannot pay for a spot and there is no secret setting. AI tools recommend businesses they can find, understand, and see other people vouching for. This guide shows you how to check where you stand in about 30 minutes, the four things that make a real difference, and the things you can safely ignore.
Are people really using AI to find businesses?
Yes, and quickly. BrightLocal’s 2026 survey of 1,002 US adults found that 45% of consumers now use AI tools to get local business recommendations, up from 6% the year before. It is one survey in one country, so treat it as a direction rather than a number to build a budget on. But the direction is hard to miss.
Two details in the same research matter more than the headline. Among people who use AI this way, 63% trust its recommendations, and 97% at least sometimes check the answer against real reviews afterwards. AI builds the shortlist, then reviews decide who gets the call. Both halves need to work for you.
Ordinary search is changing shape too. Search Engine Land’s report on a SparkToro study found that 68% of Google searches in the US ended without a click to any website between January and April 2026, up from about 60% in 2024. Where a Google AI Overview appeared, clicks fell by nearly 60%. A separate randomised trial by researchers at the Indian School of Business and Carnegie Mellon put the drop at 38%, although that paper has not been peer reviewed yet. So the fair range is somewhere between 38% and 60% fewer clicks on searches where an AI answer shows up.
There is good news if you sell locally or to people who are ready to buy. The same Search Engine Land report notes that branded searches, local business queries, and high-intent transactional searches still benefit significantly from SEO. Those are the searches most business owners care about most.
The traffic from AI tools is small but valuable. Ahrefs, an SEO software company, reported that AI search sent 0.5% of its traffic but 12.1% of its signups, with those visitors converting at 23 times the rate of organic search visitors. That is one company’s own data, so do not expect the same multiple for a plumber or a lawyer. The pattern still makes sense, though. Someone whose assistant has just recommended you arrives already half sold.
How do ChatGPT and Google AI decide who to recommend?
They behave a bit like a careful friend doing quick research. When you ask a question that needs current information, the assistant searches the web, reads a handful of pages, and writes an answer from what it finds. It tends to name businesses that show up in more than one trustworthy place and whose details match.
That gives you three things to get right. Your website has to be readable. It has to answer the question being asked. And other places on the web need to back you up.
Google is plain about the first two. Its own guidance says a page must be indexed and eligible to be shown in Search with a snippet to appear as a supporting link in AI Overviews or AI Mode. It then adds: “There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.” It also says there is no special schema markup or AI text file you need to add. So the SEO basics still apply, and anyone selling you a secret trick is selling something Google says does not exist.
The third one is where the newer evidence points. Ahrefs studied 75,000 brands and found that how often a brand is mentioned across the web tracked AI Overview visibility far more closely (a correlation of 0.664) than the number of backlinks did (0.218). Correlation is not proof of cause, and the study looked at Google’s AI Overviews only, but the direction fits common sense. Being talked about appears to matter more than being linked to, and an unlinked mention still seems to help.
How can you check if AI already recommends your business?
Before you change anything, find out where you stand. This takes about 30 minutes and costs nothing.
- Write 10 questions a customer would ask. Use real phrases from your enquiries, emails, and reviews. Mix the types: “best [your service] near me”, “how much does [your service] cost”, “who should I hire to fix [problem]”, “[your business name] reviews”, and “[you] vs [a competitor]”.
- Ask each one in four places. ChatGPT, Google (AI Mode, or a normal search that shows an AI Overview), Gemini, and Perplexity. A private browser window keeps the results less personalised to you.
- Write down what happens. For each question, note whether you are named, where in the answer, who else is named, and which websites the answer cites.
- Check what it says about you. Are your services, prices, opening times, and location right?
- Repeat every month with the same questions. Answers change from one run to the next, so ask each question twice and look for patterns, not one screenshot.
A simple scorecard is all you need:
| Question | Named? | Who else was named? | Sources cited | Anything wrong? |
|---|---|---|---|---|
| Best electrician near me | No | Two competitors | Google reviews, a directory | Not applicable |
| [Your business] reviews | Yes | Nobody | Your Google profile | Old opening hours |
You will probably land in one of three places. If you are not named at all, focus on steps 1 to 3 below. If you are named but the details are wrong, step 4 is your priority. If you are named but competitors show up more often, step 3 is usually where the gap sits.
Step 1: Make sure machines can read your website
This is the dull part, and it is where a surprising number of businesses trip up.
Check your pages are indexed. In Google Search Console, the Pages report shows which of your pages are in Google’s index and why others are not. If a service page is not indexed, it cannot be a supporting link in an AI answer. Also make sure nothing is set to noindex or nosnippet by accident, which some website plugins do.
Use real text. If your prices, services, or opening hours live inside an image or a PDF, or only appear after a script runs, an assistant may never see them. A quick test: can you select and copy the text on your page? If you cannot, fix that first.
Look at your robots.txt. Some website platforms and security plugins block bots by default. Type yourdomain.com/robots.txt into a browser and make sure you are not blocking the search crawlers of the tools you want to appear in, such as OAI-SearchBot (OpenAI), PerplexityBot, and Googlebot.
Keep the site fast and mobile friendly. That is a requirement for every kind of search, and my guide to improving your SEO has the full technical checklist. If your current site fights you on all of these, a rebuild can cost less than endless patching, and that is where web design comes in.
Step 2: Answer the exact questions your customers ask
AI tools quote specifics. A page that says “we offer competitive pricing and a tailored service” gives them nothing to quote. A page that says “a typical boiler service takes about an hour, and here is what changes the price” does.
For each of your main services, build a page or a clear section that opens with the question as a heading and answers it in the first two sentences. Then add the detail: who it is for and who it is not for, how long it takes, what affects the price, where you work, and what happens next.
Publishing prices, or at least honest ranges, is the thing most competitors will not do, which is exactly why it is worth doing. If an assistant can quote your price range and has to guess at theirs, you have the better answer.
Use your customers’ words, not yours. Pull phrases from emails, calls, and reviews, then check them against my keyword research guide. If you are new to how all of this fits together, what is SEO is the place to start, and why blog writing matters covers turning those questions into posts that keep working.
Keep the pages current. Old prices and old opening hours are how wrong AI answers start.
Step 3: Get talked about in the places AI reads
This is the slowest step and the one most owners skip, so it is also where you can pull ahead.
Reviews. Complete your Google Business Profile properly, ask every happy customer for a review with a direct link, and reply to all of them, including the bad ones. Reviews are what people check before they contact you, and they are text that AI tools can read.
The places your customers already use. Think industry directories, trade associations, marketplaces, and “best of” roundups in your sector. Ask to be listed, and check the details when you are.
Other people’s websites. Local press, podcasts, partners, suppliers, and community groups. A mention that names your business in a sentence helps, with or without a link. Offer something useful in return, such as a quote, a small piece of data, or a clear explanation of something your trade gets asked about.
Your own voice on LinkedIn and YouTube. Answer one customer question a week in your own words. Posts and short videos both give the web more places where your name sits next to your expertise. If you would rather not do that yourself, that is what social media management is for.
One warning. Never buy or fake reviews. Platforms remove them, many countries treat them as illegal, and it is not worth the risk.
Step 4: Keep the basic facts identical everywhere
AI tools blend what they find from several places. If your website says you open at 8, your Facebook page says 9, and an old directory says you have closed, the answer can come out wrong or vague.
Pick one version of the facts: business name, address or service area, phone number, opening hours, list of services, and a one sentence description of what you do. Then check it against your website, Google Business Profile, social profiles, and the main directories in your trade. Your test results are the checklist. Whatever the assistant got wrong points you to the place that needs fixing, and because Perplexity and Google show their sources, you can often trace an error back to the exact page and correct it there.
Structured data (schema markup) deserves a short note. Google says you do not need it to appear in AI features. It is still cheap to add, and it helps search engines read who you are, what you offer, and who wrote a page, as long as it matches what visitors can see. Treat it as tidy housekeeping, not a shortcut. The same goes for putting a real name and a proper bio on your articles. If you want this handled properly, it sits inside the SEO service I run for clients.
What can you safely skip?
Paid “AI ranking” packages. Nobody can guarantee a mention in ChatGPT, and if someone does, walk away. My post on whether digital marketing is legit covers the red flags to look for.
llms.txt and other special AI files. Google says you do not need them for its AI features, and I would not spend an afternoon on them before the basics above are done.
Rewriting everything for robots. Write for your customer. If a page sounds odd when you read it aloud, it is not helping anyone.
Mass produced AI content. The pages that get quoted have something specific to say. Generic copy gives an assistant nothing worth repeating.
Chasing a single screenshot. Answers change between runs and between people. Look at the pattern over a month.
Does this mean SEO and ads are finished?
No. AI answers are built from the web, so the pages that get indexed, read, and trusted are the same ones traditional SEO has always been about. What has changed is that fewer people click, so being the business that gets named matters as much as being the link that gets clicked.
Paid search still puts you in front of people who are ready to buy this week while the slower work above builds up, which is why I usually treat Google Ads and organic work as a pair. If you are still deciding where your effort should go, my digital marketing strategy guide will help you order the work.
How long will it take, and how do you measure it?
Nobody can promise a timeline. As a rule of thumb, and not a study, fixes to your own website can show up within a few weeks once pages are recrawled, while reviews and mentions build over months. Give it three to six months of steady work before you judge it.
There are four simple ways to keep score:
- Repeat your 30 minute test every month and compare the scorecard with the last one.
- Check Google Analytics for visits whose source is chatgpt.com, perplexity.ai, or gemini.google.com. The numbers will be small and will undercount, because not every click passes its source along.
- Check Google Search Console. Google says traffic from AI features is counted in your normal Performance report under Web search.
- Ask new customers how they found you, and add “AI assistant such as ChatGPT” as an option on your enquiry form. It is free, and it catches the people your analytics miss.
A 30 day plan you can start today
In week one, run the test and fill in the scorecard. In week two, fix the readability basics from step 1 and correct any wrong facts from step 4. In week three, rewrite or publish your two most important question pages from step 2. In week four, send a review request to your last 20 customers and ask for two new listings or mentions from step 3. Then run the test again and compare.
Being recommended by AI is not a new discipline. It is the old fundamentals with a new scorecard: a clear website, useful answers, and a reputation other people will back up. The businesses that do those things steadily are the ones that get named.
If you would like a second pair of eyes, book a 20-minute call and bring your test results. I will tell you what I would fix first, or whether it is worth fixing at all. If you would rather start with a wider look at where your business stands online, the free Digital Blind Spot Report is the place to begin.