Prefer a call?
GuideGEO & AEO

How to Know If AI Recommends Your Company (Free): ChatGPT, Perplexity and Google

June 24, 2026Updated November 6, 20268 min read

There are three free ways to find out whether AI recommends your company, from fastest to most thorough. First, ask ChatGPT, Perplexity and Gemini the question a customer who doesn’t know you yet would ask, without mentioning your name. Second, check whether Google’s AI answers cite you as a source. Third, read the clues you already have in Google Search Console and Analytics. That gives you an honest snapshot of today. To know whether you’re improving, repeat it every month with the same questions, or use a tool that does it for you. Here’s each step.

The mistake almost everyone makes: searching for their own name

Almost everyone’s first move is to open ChatGPT and type “what do you know about [my company]?” The answer is usually reassuring: it mentions your industry, your city, something from your site. You close the tab feeling fine.

And it tells you nothing useful. When you give it your name, you hand it the context before it answers. No new customer asks that, because they don’t know you exist yet. What all of us do is describe what we need and wait for a recommendation: “I need someone to handle commercial contracts in Texas, which firm would you recommend?”

That is the question that decides whether you show up. And it’s the one almost nobody checks.

Method 1: the question from a customer who doesn’t know you

Open ChatGPT, Perplexity, Gemini and Claude in incognito windows or with accounts that have no history. This matters: assistants know you, where you work and which companies you mention often, and from your usual account the results skew in your favor.

Write the question the way your real customer would, with their vocabulary and context. Not “what’s the best logistics company?” but something like “which company would you recommend to distribute consumer goods across Dallas with 24-hour delivery?” Ask 5 to 10 different questions: by service, by city, by problem, by type of customer.

And log everything in a table like this:

QuestionEngineDo you appear?PositionWho appearsSources cited
“Which company do you recommend for…?”ChatGPTNo—Competitor A, BDirectory X, article in outlet Y
“What options are there for…?”PerplexityYes3rdCompetitor A, B, youYour site, Google reviews

The “who appears” column is your real competition, not the one in your head. The “sources cited” column is worth even more: it maps the places where AI learns about your industry. If the same directory shows up in every answer and you’re not on it, you know where to start.

Method 2: Google’s AI answers

This one measures something else: whether Google cites you in the AI-generated box above the results (AI Overviews) and in its conversational search mode (AI Mode). Not every search triggers it. It shows up more on informational and comparison queries: “how much does… cost,” “difference between…,” “which company does… in [city].”

Run ten searches like that about your industry, also in incognito. Note how many trigger the box and whether your site appears among the cited sources. If you’re a source, you’re part of the answer. If not, you’re part of the scenery underneath, the part fewer and fewer people look at.

Method 3: the clues you already have in Search Console and Analytics

Search Console doesn’t tell you “ChatGPT cited you,” and it doesn’t separate impressions coming from Google’s AI answers either. But it leaves clues. There are three:

1

Stable position, falling CTR

Compare the last 12 months with the previous period. If your average position holds or improves but your click-through rate drops, the most likely explanation is that a generated answer resolves the question before people reach your link. I go deep on this in why your organic traffic drops even when you rank first.

2

Impressions rising without clicks rising

Google shows you more often, but the same number of people (or fewer) arrive. It’s the same signal from a different angle.

3

Visits from AI domains in Analytics

In Google Analytics 4, check your traffic sources for chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com and claude.ai. For most sites it’s still a trickle, but it’s the only direct proof that an AI recommended you and someone clicked.

No single clue is conclusive. Together, they paint a pretty clear picture.

What your results actually mean

When I go through this with companies, it almost always falls into one of three patterns:

  • AI knows nothing about you, or knows contradictory things. Your digital footprint is weak or inconsistent: your site says one thing, your Google profile another, and almost nobody else talks about you. Step one is getting your identity in order.
  • AI knows you but doesn’t recommend you. It can describe what you do, but when someone asks for a recommendation it names others. You’re missing authority: content that answers buying questions and outside sources that back you up.
  • AI recommends you, but for the wrong thing. It mentions you for services you no longer offer, in cities where you don’t operate, or with outdated details. It’s almost as damaging as not showing up, and it’s fixed by updating your information everywhere.

For all three, the full plan is in how to get recommended by ChatGPT, Perplexity and Google AI.

A snapshot isn’t monitoring

Everything above answers “do I show up today?” It doesn’t tell you whether you’re improving, which questions you’re winning, or when something changed. Models update without warning, your competitors aren’t standing still, and your customers ask in ways you never imagined. What AI says about you today may be different next month.

Tracking it by hand takes a fixed list of 20 to 50 questions, run on each engine at the same cadence, with history saved and cross-checked against Search Console. It’s doable. It’s several hours a month and a spreadsheet that ages badly. The key is keeping the list identical between measurements; change the questions and you’re comparing apples to oranges.

The rule that makes measurement worth it

Same questions, same engines, no history, every month. What changes from one month to the next is signal; what changes because you changed the question is noise.

The tools, straight up

If you’d rather not do it by hand, there are four free or low-cost options I know well because we built them:

  • The AI visibility test (aeon42.com/test-ia): free, results in a minute. You describe your company and it asks an AI assistant with web search the question a customer who doesn’t know you would ask. It tells you what AI knows about you, who it recommends instead and three actions. It’s a one-moment snapshot with a single assistant, not monitoring, and taking it subscribes you to the newsletter.
  • The free tools at aeon42.com/herramientas: they check whether ChatGPT, Claude and Perplexity crawlers can get into your site, inspect your structured data and detect whether Google’s AI answers cite you.
  • Compare your site (our free tool): it doesn’t ask AI anything, but it checks whether crawlers can get in and read your homepage, and tells you which percentile you land in compared with Mexico’s largest companies. It’s the first filter: if AI can’t read you, it won’t recommend you.
  • The full platform: AEON42 connects to your Search Console, analyzes 16 months of your data and measures your AI visibility every week. The Solo plan covers Google’s AI answers and AI Mode; the Studio and Agency plans add ChatGPT, Gemini, Perplexity and Claude. From USD $49 a month, with a 14-day trial.

And if what you need isn’t measuring but fixing, with a team that diagnoses, implements and delivers results, that’s what our Google and AI search positioning service is for.

Frequently asked questions

For a quick, free snapshot, the AI visibility test at aeon42.com/test-ia asks an assistant with web search the question a customer who doesn’t know you would ask, and tells you whether you appear and who it recommends instead. For weekly tracking across several engines you need a platform like AEON42, or a fixed list of questions you run yourself every month in ChatGPT, Perplexity, Gemini and Claude.

Yes. Open ChatGPT in incognito or with an account that has no history, ask the questions your customer would ask without mentioning your name, and note whether you appear, in what position and who appears instead. Repeat the same in Perplexity, Gemini and Claude.

Not directly: it doesn’t separate impressions coming from AI Overviews or AI Mode. But it leaves clues: a CTR that falls while your position holds, and impressions that rise without clicks rising. For direct proof, check Google Analytics for visits from chatgpt.com, perplexity.ai and other AI domains.

Define a fixed list of 10 to 15 commercial questions in your industry and run them every month on the same engines. Record which brands appear, in what order and with which sources. If a competitor starts appearing where you used to be, look at what they published and who cited them: that’s the signal the model picked up.

Once a month is enough for most companies, always with the same questions so the comparison holds. If you’re making major changes to your site or content, check every two weeks during that period.

Ready to implement it?

Free diagnosis of your search and AI visibility

We analyze your site, your visibility on ChatGPT and Perplexity, and deliver a prioritized roadmap. No strings attached.