How to check whether AI recommends your business

Most business owners run this test once. They type their own name into ChatGPT, see something reasonable, and move on. That test tells you almost nothing, because your customers aren’t asking for you by name. They’re asking who to hire.

Here’s a version worth running.

1. Write the questions your customers ask

Skip your brand name. Write ten questions the way a customer would type them in a hurry, lowercase and a little messy:

• “who’s the best [what you do] in [your city]”
• “[problem you solve] near me”
• “is [your service] worth it for [type of customer]”
• “[your service] vs [the alternative people compare it to]”

Mix the ones that name a place with the ones that don’t. Mix buying questions with research questions. Ten is enough to start.

2. Ask more than one engine

Your customers don’t all use the same tool. Run every question through at least these:

Engine	Why it matters
ChatGPT	The most-used chatbot, with web search built in
Google AI Overviews and AI Mode	Sits above the results most people already use
Perplexity	Shows its sources, which tells you what it trusts
Gemini and Copilot	Reach through phones, browsers and work tools

Use a private window, stay logged out where you can, and note the location the engine thinks you’re in. Location changes local answers a lot.

3. Ask each question more than once

AI answers aren’t fixed. The same question can name different businesses on different runs. So ask each one three times, on different days if you can. One mention in one run is noise. Three out of three is a position.

4. Record what you see, not what you hoped to see

For every run, write down four things:

1. Named or not. Did the answer mention your business?
2. Position. First, second, third, or buried in a list?
3. What it said. Is the description of you accurate?
4. Sources. Which websites did the engine cite? Perplexity and Google show these. They’re a map of what the engine trusts in your market.

The fourth column is the most useful. If every answer cites the same directory, the same review site or the same competitor’s blog, you now know where the engines are getting their picture of your market.

5. Read the result honestly

Count it up. If you were named in 4 of 30 runs, your baseline is 4 of 30. Don’t round it up to “sometimes.” And if it’s zero, write zero. An honest baseline is the only thing you can measure progress against.

Then look at who was named instead. What do they have that you don’t? It’s usually one of four things: a site that clearly says what they do and where, consistent listings, a steady stream of recent reviews, or pages that answer the exact question asked.

6. Run it again next month

Same questions, same engines, same method. One snapshot doesn’t tell you much. What you want is a number you can track over time, checked on dates you set.

That’s the loop AuthEngine.ai runs for clients, at a bigger scale and on a schedule. Read why it matters, or tell us what you do and we’ll tell you what we’d measure first.