AI assistants

What ChatGPT says about you and what you can do about it

Where ChatGPT, Gemini and Perplexity get their information about people, why they sometimes get it wrong and how to make the answer right.

Written by
Eduard
Published
Reading time
2 min
A laptop screen open in a dark room, filled with lines of colorful code, with the keyboard lit by the screen.

More and more people no longer stop at Google. They ask ChatGPT, Gemini or Perplexity directly: “Who is this person?” or “What does this company do?” The answer comes back as a short paragraph, phrased with confidence, and the person asking often takes it at face value.

That paragraph has become a first impression you don’t write. It’s worth knowing what it says and where it comes from.

Where the information comes from

AI assistants don’t know you. They answer based on texts they have read or that they find at the moment of the question. A few sources come up often:

  • Websites. Your personal website, your company website, profile pages.
  • The online press. Interviews, articles and mentions in online publications.
  • Public profiles. Social media, LinkedIn, company pages.
  • Other pages that mention you. Events, lists, reviews, podcast transcripts.

Some assistants search the internet before answering and show their sources. Others rely more on what they learned earlier. Either way, what counts is what exists publicly about you and how clearly it is phrased.

Why they get it wrong

When information is scarce, the assistant has too few reference points. It may confuse two people with the same name, pick up an old job title or fill the gaps with guesses that sound plausible.

Errors also appear when sources contradict each other: one site says one thing, another says something else. Without a clear, current source, the assistant has no way of knowing which version is right. For your reputation, inconsistent information is almost as risky as false information.

What you can do

Start with an audit. Ask several assistants the same question and note what they answer, what sources they cite and what’s missing. Repeat the check regularly, because the answers change.

Then build a reference source about yourself. A personal website with a complete biography, with your name spelled identically everywhere, your current role and your company, helps both people and AI assistants understand who you are.

Third, make sure you are present in credible places, with consistent information: interviews, articles, podcasts, up-to-date profiles. The more your story is told the same way across trusted sources, the harder it is to misunderstand.

What you can’t control

You can’t write an AI assistant’s answer directly, and nobody can guarantee what it will say. Anyone who promises you that is promising something they can’t deliver. What you can do is make sure the correct information about you is easy to find and easy to understand. The rest depends on how each assistant works.

My recommendation

Treat AI assistants’ answers like any other reputation channel: you monitor them and work on the sources, not the symptom. That is what PRAI Academy does for its clients, by regularly checking the answers and working on the public sources they are formed from: your personal website, the online press and your profiles. If you haven’t yet checked what is being said about you, that is where I would start.

  • AI assistants
  • ChatGPT
  • online reputation
Keep reading

More notes on image, the online press and AI assistants.

Next step

Your first conversation, straight with the strategists.

We clarify what you want to convey and the package that gets you there. You sign the contract; we take care of the rest.