How to check if ChatGPT recommends your brand

It sounds trivial. But that simple test sits on top of a shift most businesses haven’t noticed yet.

More and more people don’t type questions into Google. They ask ChatGPT, Perplexity or Gemini and get one finished answer: a specific company, product or service. Not ten links to sift through. One recommendation. If that answer names a competitor instead of you, you lose the customer, and most of the time you never even know.

In search you can see when you slip off page one. In an AI answer there is no “page two”. You’re either in it or you’re not.

Key takeaways

• AI gives one answer, not a list of links. Either your name comes up or a competitor’s does; there’s no “bottom of page one”.
• A manual test takes minutes and anyone can run it. Ask ChatGPT the question your customer would ask and see if you’re in the answer.
• One test isn’t enough. AI answers shift with every prompt, so you only get a real signal after several repeats.
• Every assistant sees the web differently. You can be strong in Gemini and absent from ChatGPT.
• Ranking in Google guarantees nothing. AI often cites sources other than the top of the search results.
• AI recommends the best-described source, not the best product. That’s bad news and an opportunity at once, because it’s something you can influence.

Why does it matter whether AI recommends your brand?

Picture a customer looking for a service in your field. A year ago they’d have typed a phrase into Google and skimmed a few offers. Today they increasingly open ChatGPT and ask outright: “what do you recommend?”

They get an answer with two, maybe three names. The rest of the market simply doesn’t exist for that customer, in that moment.

That rewrites the rules. In Google, your position on the list mattered. In AI, what matters is being in the answer at all. There’s no consolation prize for “fourth place”: either your name comes up or it doesn’t. And because the model decides who to mention, you’ve lost control of the first impression: it’s now shaped by how the model pieced your industry together from what it found online. So the first step isn’t optimization. It’s measurement.

How can I check if ChatGPT recommends my brand myself?

The good news: you can run the basic test yourself, in a few minutes, for free. Open ChatGPT with web search on and step into the shoes of a customer who doesn’t know you yet.

The key rule: don’t ask “do you know company X?”. The model will politely confirm almost anything. Ask the way someone looking for a solution asks, not someone looking for a specific brand. Try a few question types:

• Category discovery: “Which companies do you recommend for [service]? Name specific ones.”
• Buying intent: “I’m looking for [product] for [type of customer]. What do you recommend and why?”
• Comparison: “Compare [your brand] with [competitor]. Which should I pick when [scenario]?”

Note whether your name appears, in what position and in what context, and repeat each question a few times in fresh sessions. A manual test answers “am I here today?”, but it doesn’t answer “is this better than a month ago?”. Treat one run as a starting point, not a verdict.

Why do the answers keep changing?

Run the test in the morning and you’ll get one set of companies. Repeat it in the evening and you’ll get another. That’s not a bug; it’s how these models work.

Assistants have randomness built in. The same question returns a slightly different answer each time, in a different order and sometimes with different names. On top of that the sources shift: what AI cites this week it may swap out next.

The consequence is simple but important. One test proves nothing. A sensible picture only emerges from an average, several repeats of the same question, watched over time. You don’t care about a single answer. You care about the share of answers you show up in, and whether that share is rising.

How is visibility in ChatGPT, Perplexity and Gemini different?

The biggest mistake is assuming there’s one “AI visibility”. There isn’t. Each assistant builds its own picture of the web, drawing on different sources and weighing them differently. You can dominate one engine and be completely absent from another, with the same site and the same content.

So run the test in parallel across ChatGPT with search, Perplexity, Gemini and Google AI Overviews (the AI answer that now sits at the top of a normal Google search), using the same set of questions. Only side by side do you see where your real gap is, and which surface matters most for you.

From my own experience

I measured this on my own product. For remote-work questions, Gemini recommended my site in most answers. ChatGPT recommended it in none. Same site, same content, same questions. Had I checked only one of them, I’d have drawn a completely wrong conclusion: either “this is great” or “this is hopeless”. The truth sat somewhere in between and differed engine by engine.

Piotr Czerwiński · Founder, CiteLyzer & HiddenJobs

What should I do if AI doesn’t recommend me?

Say the test went badly. First understand one mechanism: AI doesn’t recommend the best product. It recommends the best-described source it found. A competitor doesn’t need a better service; they may just have a better article, ranking or review that AI ran into. That’s bad news for the ego and good news for strategy, because it’s something you can influence. A few concrete directions:

• Earn mentions beyond your own site: industry rankings, directories, reviews, quotes in articles. Where and how often the web talks about you weighs more for AI than raw link count.
• Publish content that answers customer questions directly, not “about us”, but “how to choose X”, “X vs Y”, “how much does Z cost”.
• Tidy up your on-site information with structured data (schema for your organization, services, reviews, FAQ) so AI understands what you do.
• Check where AI knows a competitor from. If it cites a specific ranking you’re missing from, you have a ready-made target.

And the most important part: measure again after every change. AI visibility isn’t a one-off task; it’s a loop. You make a change, wait, check the trend, adjust. Without that loop you can’t tell what works from what only looks like a good idea. Doing this by hand works to start, but it quickly turns into a full-time job, which is the gap CiteLyzer is built to close: one view of where you show up across ChatGPT, Perplexity, Gemini and Google AI Overviews, and how it moves over time.