Istvan runs a projection-mapping studio in Budapest. A new client in the United States got in touch. Asked how they had found the studio, the client said they had asked ChatGPT: who is the best immersive experience maker? It named Istvan’s studio.
Not because of their SEO. Because of what they had actually done: a Guinness World Record and twenty years of real work, written about in places the AI had read. That’s hard to fake, and hard to push. Which is exactly why it works.
The prompt
Ask the question your customer would ask. Put in what you do and where:
The “why those?” matters most. It tells you what the AI counts as proof in your field.
Then ask about yourself
If the answer is thin or wrong, that’s your to-do list. Ask:
How to read the answers
- Ask more than once, and in more than one app. ChatGPT, Claude, Gemini and Perplexity answer differently, and the same app can answer differently tomorrow. One answer is a sample, not a ranking.
- Check every claim it makes about you. It can be confidently wrong. If it is, the fix is a clear, true page on your own site, not an argument with the chat.
- Look at who it names instead. Then look at what they have that you don’t: awards, press, published work, reviews, a page that says plainly what they do and for whom.
What actually moves it
The same things that make a person trust you: real work, shown in public. Say on your own website, in plain words, what you do, where, and what you have achieved. Put the proof where others can see it: awards, press, client stories, talks. Then keep doing the work worth writing about.
There’s no prompt that makes an AI recommend you. There is one that tells you what it thinks, and that’s the one to run first.
