What a 20-Search Study Found
When ChatGPT recommends a local contractor, it builds a list of businesses first, then reads from the top. The first business on that list is the one it names in writing, almost every time.
LinkTech Solutions ran a 20-search study on September 27, 2026, asking ChatGPT (with web search enabled) to recommend 20 local businesses across Orange County and the Inland Empire. In 18 of 19 answers, the first business on ChatGPT's own list was the first one it named in its written answer. In all 19 answers, every business it recommended from the list appeared in the same order it was listed, with no position skipped.
The question "how do I get ChatGPT to recommend my business" is, for local searches, mostly the question "how do I get to the top of the list it builds first." The written answer adds details like license numbers and a line about each business, but it rarely changes who gets named.
ChatGPT also named businesses that were not on its own list in 4 of 19 answers, all Inland Empire queries. Those names came from the directories it cited rather than from its internal list. For a deeper breakdown of the full pipeline, see how ChatGPT picks local businesses.
Directories, Not Websites, Feed the List
Of 53 sources ChatGPT cited across the 20 searches, 45 were directories or review aggregators like Birdeye, Expertise.com, and Angi. Six were State Bar or county bar pages. Only two were business websites.
This matches broader findings from Yext, which analyzed 6.8 million AI citations and found 42% came from listings and 44% from business websites. For local service queries specifically, the directory share runs higher. ChatGPT builds its list from the directories it trusts in a given trade and city, then pulls details from those listings to write the answer.
A contractor missing from the directories that matter in their trade, or listed with a different name, phone number, or category, gives ChatGPT nothing to quote. For a full breakdown of which directories ChatGPT actually reads, see which directories ChatGPT reads when recommending contractors.
Review Count Does Not Set the Order
The most-reviewed business was first on ChatGPT's list only twice in 19 answers, according to the LinkTech study. Review volume influenced the list, but it did not control the top position.
About 10% of businesses on the lists had fewer than 10 reviews, and roughly 13% sat below 4.5 stars. Some businesses ChatGPT recommended had ratings as low as 3.2. The review bar is lower than most contractors assume, and it varies by trade. In some trades, a few dozen reviews puts a business in the top three.
SOCi's 2026 Local Visibility Index, which analyzed 350,000 business locations across 2,751 brands, found ChatGPT recommends only 1.2% of them. Perplexity recommends 7.4%, Gemini 11.0%, and Google's local 3-pack includes 35.9%. The gap between ranking on Google and getting named by ChatGPT is the gap between 36% and 1.2%.
Location in the City Beats Review Count
Out-of-town businesses with far more reviews ranked below local businesses in the LinkTech study. Being clearly located in the city the question asked about mattered more than having the most reviews.
This is a specific problem for contractors who serve multiple towns from one shop. A plumber in Clifton serving Bergen County whose listing and website say "Clifton" while the customer asks "plumber in Hackensack" may not land at the top. A real address or a stated service area in the target city, matching across the listing and the website, keeps the business in the right list.
License and Trade Records Act as Verification
In the 4 answers where ChatGPT named businesses not on its own list, those names came from the directories it cited. The assistant used State Bar records and county licensing pages to verify them.
A contractor whose license record, BBB listing, or trade association membership matches their directory listing gives ChatGPT a second source that agrees. When details disagree across sources, the assistant loses confidence the business is real and current, and it skips them for a safer answer.
What Contractors Should Do
- Find the directories your trade is cited from. Ask ChatGPT to recommend a contractor in your city and note which directories appear in the answer. Claim or correct your listing on each one.
- Keep name, address, phone, and category identical everywhere. One inconsistent phone number across directories is enough for ChatGPT to pass you over.
- Be clearly in the city you want to be recommended in. A real address or a stated service area, the same on your listing and your website.
- Get onto the license and trade records ChatGPT uses to verify. Keep your license record, BBB listing, or association membership consistent with your directory listing.
- Keep a website that says plainly who, what, and where. Most listed businesses had one, and the listings link to it. For B2B queries with fewer directories, the website is often the source.
- Check the answer every month. Ask ChatGPT the question your customers ask, with your city in it, and write down who gets named. That is the only way to know whether the work is paying off.
FAQ
How does ChatGPT choose which local businesses to recommend?
ChatGPT builds a list of businesses from directories and review aggregators, then writes its recommendation from the top of that list. In a 20-search study by LinkTech Solutions (September 2026), the first business on the list was the first one named in writing 18 of 19 times. The order leaned on location in the city and review volume, but the most-reviewed business was first only twice.
Does my website matter for ChatGPT local recommendations?
For local service queries, ChatGPT relies more on directories than business websites. In the LinkTech study, 45 of 53 cited sources were directories and only 2 were business websites. A clear website still matters because directories link to it, and for B2B queries with fewer directories, the website is often the source ChatGPT cites.
How many reviews does a contractor need to get recommended by ChatGPT?
The threshold varies by trade. In the LinkTech study, about 10% of listed businesses had fewer than 10 reviews, and some recommended businesses had as few as 3 to 9 reviews. Review count mattered less than list position and location in the city being searched.
What star rating does ChatGPT require to recommend a business?
There is no hard threshold. In the LinkTech study, 13.2% of listed businesses sat below 4.5 stars, and some recommended businesses had ratings between 3.2 and 3.8. SOCi's 2026 data shows businesses averaging 4.3 or higher across multiple platforms tend to appear more often, but a lower rating does not automatically exclude a business.
How do I check whether ChatGPT recommends my contracting business?
Ask ChatGPT, with web search on, the question your customers ask. Use your city and trade in the question. Write down who gets named and whether the details are correct. Repeat the same question monthly to track whether your position changes.
Run your free AI visibility audit to see where your business stands across ChatGPT, Gemini, Perplexity, and Google AI Overviews.