To benchmark your AI visibility against competitors, run one identical prompt panel for the whole competitor set, on the same providers, on the same day, and lay the results out as a matrix: prompts down the side, brands across the top. Same-day execution is not a nicety; AI answers drift constantly, so a comparison assembled across different weeks is comparing two different weathers, not two brands.
The method: same prompts, same providers, same day
Pick your three to five most direct competitors and build a panel of organic, unbranded buyer prompts: the recommendation, comparison, alternatives, and pricing questions a buyer in your category actually asks. Do not seed the prompts with any brand name except where the prompt genuinely is branded ("alternatives to [leader]"). Run the full panel in one session per provider, several runs per prompt, and record every brand mentioned in every response with its position. One pass produces the whole competitive picture, because each response is itself a head-to-head result: whoever appears, appears against everyone who does not.
Reading the matrix
The matrix makes three patterns jump out that no summary score can show. Column patterns: a competitor with a filled column across comparison prompts has comparison content the engines are retrieving, and you can go read exactly what it is. Row patterns: a prompt where nobody in your set appears is open territory, usually because the engines are citing generic publishers instead of vendors. Block patterns: strength in one provider and absence in another, which is common because Perplexity retrieves live sources while ChatGPT leans on training data, so the same brand profiles differently on each.
What gaps mean by prompt type
- Losing recommendation prompts: you are absent from the listicles and best-of coverage the engines draw on. An off-site content and PR problem.
- Losing comparison prompts: no vs content exists, yours or third party, that includes you. The most directly fixable gap.
- Losing alternatives prompts: nothing on the retrievable web frames you as an alternative to the category leader. A positioning problem more than a content volume problem.
- Losing pricing and fit prompts: your pricing is unpublished, or published in JavaScript that AI crawlers cannot read.
Finding the prompts competitors rank for
You cannot look up a competitor's winning prompts directly, because no AI provider exposes prompt logs for anyone. The benchmark panel is how you reconstruct it: every response that names them and not you is, by definition, a prompt they rank for, and on cited surfaces like Perplexity you can also see which page earned them the spot. Reading your own side of the matrix is covered in which prompts your brand ranks for. If you want the matrix without building the machinery, the free scan produces a competitor visibility matrix across five providers as standard output.
See the full competitive matrix in one scan
The free scan runs identical buyer prompts across five AI engines and returns a competitor visibility matrix, so you can see exactly which prompts you win, lose, and have not contested.
Frequently asked questions
How do I audit a competitor's AI visibility?
Run the same unbranded buyer prompts you use for your own brand and record every response that names them: which prompts, which providers, what position, and on cited surfaces like Perplexity, which source pages earned the mention. Since no provider publishes prompt logs, this sampled reconstruction is the only audit possible, for them and for you.
What software benchmarks AI visibility against competitors?
AI visibility platforms such as Profound, Peec, AthenaHQ, Otterly, Scrunch, and YouGotRanked track competitor mentions across providers. Whatever you pick, verify it applies identical prompts and same-day methodology across all brands; a benchmark built from different prompts or dates is not a benchmark. YouGotRanked's free scan includes a competitor visibility matrix across five providers.
How many competitors should I benchmark against?
Three to five direct competitors keeps the matrix readable and the work sustainable. Include the category leader even if you never win deals against them, because alternatives-to-leader prompts are among the highest intent queries in AI, plus the rivals who actually appear in your lost deals.
Related answers