How to check if ChatGPT recommends your business

To check if ChatGPT recommends your business, ask it the questions your customers ask, by service and town, run each one several times in a fresh chat, and count how often you’re named. Then do the same in Perplexity and Google’s AI Overviews, since each builds its answer a different way. One run tells you very little. Five runs of a handful of prompts on each engine tell you where you stand.

This is the same check we run for clients every month. Here is how to do it yourself, with prompts you can copy, the reason a single answer can’t be trusted, and a table to log it in.

Build your prompts from service and town

Write the prompts the way a customer types them, not the way you’d describe yourself. Customers ask for a service in a place. They don’t ask for your business by name, so leave your name out: asking “is [your business] any good?” tests whether the engine has heard of you, not whether it would send someone your way.

Start from this template and swap in your service and the towns you serve:

• best [service] in [town], Utah Copy
• who should I call for [service] in [town]? Copy
• [service] near [town] with good reviews Copy
• top rated [service] in Cache Valley Copy
• I need a [service] in [town] this week. Who do you recommend? Copy

Filled in for a few Cache Valley trades, that looks like:

• best pest control in Logan, Utah Copy
• who should I call for a clogged drain in Smithfield? Copy
• house cleaning near North Logan with good reviews Copy
• top rated roofer in Cache Valley Copy
• I need a dog groomer in Hyde Park this week. Who do you recommend? Copy

Five to ten prompts is plenty. Cover your main service in every town you want work from, then add one or two for your side services. If most of your jobs come from Providence, Providence belongs in the list even though Logan is the bigger town.

Run each prompt the same way every time

Consistency matters more than any single setting. In our checks we open a fresh chat for every run, so the answer isn’t shaped by the conversation before it, and we use a private window or a signed-out session where the engine allows it, so saved memory and past searches stay out of it. The town goes in the prompt itself rather than relying on the engine to guess your location.

For each run, write down every business the answer names, not just whether you made it. The list of names is the useful part. It shows who the engine trusts for your service, and that’s the list you’re trying to get onto.

Why one run is not enough

AI answers are generated fresh each time, so the same question gets a different list on the next try. SparkToro, a marketing research company, measured how much. In its study, 600 volunteers ran 12 prompts through ChatGPT, Claude, and Google’s AI a combined 2,961 times. The finding: there’s less than a 1 in 100 chance that ChatGPT or Google’s AI gives the same list of brands in any two responses, and less than 1 in 1,000 that two lists come back in the same order.

That kills the idea of an AI “ranking.” Being third in one answer means nothing when the next answer reshuffles the list. SparkToro’s founder, Rand Fishkin, came out of the study calling “visibility % across dozens to hundreds of prompts run multiple times” a reasonable metric, and said any tool claiming a “ranking position in AI” is “full of baloney.”

So count mentions. We run each prompt at least three to five times per engine and record the share of runs that name the business. Named in four of five runs is a solid recommendation. Named once in five means you’re on the edge, and it can swing either way next month.

How ChatGPT, Perplexity, and Google AI Overviews differ

The three engines don’t pull answers from the same place or show their work the same way, so check all three.

ChatGPT. OpenAI says its OAI-SearchBot crawler “is used to surface websites in search results in ChatGPT’s search features,” so the answers you’re checking come from what that crawler has read. If an answer includes links to the sites it drew on, note which of your listings they point to.

Perplexity. Perplexity’s crawler docs say that when someone asks a question, Perplexity “might visit a web page to help provide an accurate answer and include a link to the page in its response.” Its source list makes Perplexity an easy place to see where your information comes from.

Google AI Overviews. Google says AI Overviews and AI Mode may run extra related searches, a “query fan-out,” and can show “a wider and more diverse set of helpful links” than a classic search. An Overview doesn’t appear on every search, so a missing Overview on a prompt isn’t the same as being left out of one.

Google is the one engine that reports this for free. Sites that show up in AI Overviews and AI Mode are “included in the overall search traffic in Search Console,” in the Performance report under the Web search type. Since June 2026, Search Console also has a separate Search Generative AI performance report, and Google says it had “rolled out these insights to all websites worldwide” as of August 31, 2026. The launch post lists impressions, pages, countries, devices, and dates. It doesn’t list clicks, so treat it as a count of how often you appeared, not how many people came through.

A monthly tracking table

Keep it plain. One row per prompt per engine, filled in on the same day each month:

The numbers in those rows are made up to show the format. Your first month is the baseline. After that, read across the months: is “times named” climbing on each engine, and is the “named instead” column shrinking or changing? A competitor who shows up on every engine for every prompt is the business the engines trust most for your service, and that’s the name you’re trying to sit next to.

Put your Search Console generative AI impressions next to the table each month. Hand-run prompts tell you about the questions you chose. Search Console tells you about the questions real people asked.

What to do if you are not named

Not being named is the normal starting point, not a sign something is broken. SOCi, a local marketing company, found in its 2026 Local Visibility Index that ChatGPT recommends only 1.2% of business locations and Perplexity 7.4%, against 35.9% of those same brands showing in Google’s local 3-pack. That study covered brands with 50 or more locations, not single-shop businesses, but the gap it shows is the same one most local owners find on their first check.

The check tells you where you stand. It doesn’t tell you why the engines picked someone else, and the reasons differ from one business and one town to the next. Sorting that out is the part we do for clients: we look at what the engines read about you, find where it falls short, and fix it. Then the same check, run the same way next month, shows whether it worked.

Have us run it for you

Our Managed SEO add-on is $300 a month on top of any Vient website plan. It includes this check every month across ChatGPT, Perplexity, and Google, along with rank tracking and a monthly report, plus the work that gets you named. The full list is on our SEO services page, and every plan is on the pricing page.