AI SEO
AI Visibility Prompt Builder
Build the prompt set that tells you whether ChatGPT, Gemini and Perplexity recommend you or your competitors.
Run each one in ChatGPT, Gemini, Perplexity and Claude, then record whether is named and what it's named alongside. Re-run the same list monthly — the movement is the measurement.
In short
You can't measure AI visibility until you've decided which questions count. This builds a prompt set across the three stages a buyer moves through — problem, comparison and decision — so you can run the same list every month in ChatGPT, Gemini, Perplexity and Claude, and see whether assistants name you or name your competitors.
Why A Prompt Set, Not A Score
Plenty of tools will sell you an AI visibility score. The number moves, and you learn nothing about why, because a single figure hides the thing you can act on: which questions you're absent from.
A prompt set is the opposite. Thirty questions, run monthly, recorded honestly. When you drop out of "best link building companies" but hold your position on "is Degen any good", you know exactly what to fix.
The Three Stages
Problem. They haven't named a solution yet — they're asking what to do about something. You almost never get named here, and that's normal. These prompts tell you which publications and sources the assistants trust in your category, which is your digital PR target list.
Comparison. They're building a shortlist. This is the stage that matters most and the one worth obsessing over, because being on the list is the whole game. If assistants name five companies and you're not one of them, nothing downstream helps.
Decision. They've got your name and they're checking it. Losing here usually means a review problem or a thin brand presence rather than a search problem.
How To Run It
- Generate your set and copy it.
- Run every prompt in ChatGPT, Gemini, Perplexity and Claude. Use a fresh chat each time, with memory and personalisation off, or you're measuring your own history.
- Record three things per answer: were you named, who else was named, and what was cited.
- Repeat monthly. The movement is the measurement — a single snapshot tells you almost nothing.
What To Do With The Answers
The citations are the most useful column. Assistants pull from a fairly small set of sources per topic, and they're usually the same ones each month. When the same three publications keep coming up and you're not in any of them, you've just found your outreach list — and getting mentioned there does more for AI visibility than anything you can change on your own site.
That's the uncomfortable part of AI search: most of the work happens off your domain.
FAQ
Frequently asked questions
How Do I Check if ChatGPT Mentions My Brand?
Ask it the questions your buyers ask, not questions about your brand. Start a fresh chat with memory turned off, run a set of category and comparison prompts, and record whether you're named and who's named alongside you.
What Is AI Visibility?
How often AI assistants name your brand when someone asks a question you should be the answer to. It's separate from search rankings — assistants pull from citations, mentions and third-party sources as much as from your own pages, so you can rank well and still be invisible.
How Often Should I Track AI Visibility?
Monthly is enough for most brands. Answers vary between runs even with identical prompts, so a single check tells you very little — the trend across several months is the signal.
Why Do I Rank on Google but Not Get Mentioned by AI?
Assistants weigh what other sites say about you more heavily than what you say about yourself. If your visibility comes from your own pages and you've got few independent mentions, reviews or citations, there's not much for an assistant to draw on.