How to Check What AI Says About You (Simple Methods)

Checking your own reputation with an AI tool is less like checking a credit score and more like eavesdropping on a conversation about you at a party you weren't invited to. It takes a few minutes, costs nothing, and it's the single most useful thing you can do before deciding what, if anything, to fix.

Why this is worth doing at all

People increasingly ask AI tools the kinds of questions they used to type into a search bar: who's good at this, what's true about that, should I trust this person or company. If you've never actually looked at what comes back for your own name or your own field, you're making decisions about your visibility based on a guess. A five-minute check turns that guess into an actual, specific answer you can act on.

Method one: ask your name directly

This is the obvious first move, and it's useful, though it's not the whole story. Type your full name into ChatGPT, Gemini, and Perplexity and see what each one says. Note whether the answer is accurate, vague, confidently wrong, or a flat "I don't have information about that." Try a couple of variations too, your name with your city added, your name with your job title added, since sometimes a little context changes what the model can retrieve.

Method two: ask the question a stranger would actually type

This is the more revealing test, and the one people skip. Instead of asking about yourself, ask the question a real prospective client or contact would type when they don't know you exist yet: "who's a good [your field] near [your area]," or "who should I talk to about [the problem you solve]." This shows you whether you'd actually get found in the moment that matters, rather than just whether the model recognizes your name when prompted directly. The full method for phrasing this properly and reading the results is covered in does ChatGPT recommend you.

Method three: check for the AI Overview on a regular search

Beyond the chatbots, plain Google searches now often show an AI-written summary above the regular results, commonly called an AI Overview. Search your own name, or the buyer-style question above, in regular Google, and see whether that summary box appears, and if it does, whether it mentions you. This is a different surface than ChatGPT, built from different mechanics, but it matters just as much since a huge volume of everyday searches pass through it. For how these overviews actually get built, Google AI Overviews explained is worth reading alongside this check.

Method four: check the spelling and nickname variants

If you go by more than one version of your name, a shortened first name, a maiden name, a name with or without a middle initial, run the check under each variant. It's common to find that one spelling returns a solid answer while another comes back empty, which tells you exactly where your online trail is thin or fragmented.

Method five: compare yourself to a peer, gently

Ask the same buyer-style question with a well-known peer's name swapped in, or with no name at all, just the general question. Seeing what a stronger or more established peer generates for the identical question gives you a useful sense of what a fuller answer actually looks like, so you're not just staring at your own blank result without context for what "good" would look like.

Figure: run all five before drawing any conclusion about where you stand.

Mistakes that quietly ruin the check

The biggest one is changing your wording between checks, which makes it impossible to tell whether a different answer reflects real change or just a different question. Another is treating a single check as final proof of anything, when these systems can and do vary their answers slightly even for an identical question on different days. A third is only checking the one tool that happens to already know you and skipping the others, which feels reassuring but hides the real gaps. And a fourth is forgetting to screenshot or write down what you find, so a month from now you have no baseline to compare against.

What to do with an honest but disappointing result

If nothing comes back for you anywhere, that's a completely normal starting point, not a verdict on your career, and it points you toward publishing and earning mentions rather than toward panic. If something comes back but it's wrong, outdated, or attached to the wrong person, that calls for a different, more targeted response, which what if AI says something false about me walks through step by step. And if you're mentioned but only faintly, buried at the bottom of a longer list, that usually means your foundational signals exist but aren't dense enough yet to win the tiebreak against better-documented competitors.

Turning this into a habit, not a one-time panic check

Run this same set of checks again in a month, with the exact same questions, and keep a simple running note of what changed. A single check is a snapshot. A repeated check, done the same way each time, is the only thing that actually tells you whether your efforts are working. Set yourself a calendar reminder for it, because this is exactly the kind of useful habit that quietly slides once the initial curiosity wears off.

If you're checking on behalf of a business or a team

Everything above works just as well for a company name as it does for a personal one, with one extra wrinkle worth knowing. Ask about the business by name, and also ask the buyer-style question without naming the business at all, since that shows you whether a prospective customer who has never heard of you would actually be pointed your way. It's also worth checking a few individual team members by name if they're client-facing, since a company can be well documented while its actual people remain invisible, or the reverse, a well-known founder while the company name itself returns nothing useful.

Keeping a log you can actually act on

A simple running note works better than trying to remember results in your head. For each check, record the date, the exact tool, the exact question you typed, whether you appeared and where in the answer, and any reasoning the tool gave. Over a few months this turns from a curiosity into a genuinely useful record, showing you which engine is catching up to your published work and which one still needs more to go on. It also protects you from the natural but misleading urge to judge your progress off a single lucky or unlucky answer.