How to Measure AI Visibility When There's No 'Rank #1'

In AI search there's no 'position 3' to celebrate — either the answer names you or it doesn't. So measuring AI visibility means tracking four things instead of one keyword rank: whether you're cited, how often, against which competitors, and whether the engines can even reach your pages. Set a baseline now and the work becomes provable rather than anecdotal.

Build a prompt library — 10 to 20 questions a real buyer would ask in your category, branded and unbranded — and run them across ChatGPT, Gemini, Perplexity, Copilot and Google's AI Mode. For each, log the outcome: cited (named with a source), mentioned (named, no link), or absent — and note which competitor showed up instead. Re-run monthly; the trend is the signal.

Layer in the platforms you don't control. Track whether your domain appears in Google's AI Overviews for core terms, and check crawl evidence — Google Search Console's crawl stats plus your server logs for GPTBot, PerplexityBot, ClaudeBot and friends. If the bots aren't fetching you, nothing downstream can.

None of this needs expensive tooling to start — a spreadsheet and a disciplined monthly cadence beat a dashboard you never check. The point is before-and-after honesty: know where you stand today, fix the biggest gap, and watch the citations move. That's the whole game.