How do I spot the topics where competitors outrank us in AI answers?

The core method is to build a prompt set, run it across AI engines repeatedly, and compare who gets mentioned or cited for each topic.The core method is to build a prompt set, run it across AI engines repeatedly, and compare who gets mentioned or cited for each topic.

1. Build a prompt set that mirrors real questions Pull from sources that reflect how your buyers actually ask: sales call transcripts, support tickets, search console queries (especially long, question-style ones), "People also ask," and Reddit or community threads. Group the prompts into topic clusters (e.g., "pricing comparisons," "how to do X," "best tool for Y," "alternatives to Z"). Aim for a few dozen prompts per cluster, and include a mix of unbranded, comparison, and "alternatives" queries.
2. Run them across the engines your audience uses That typically means ChatGPT, Google AI Overviews/AI Mode, Perplexity, Gemini, and Copilot. Answers vary run to run, so repeat each prompt several times (say 5-10) and track rates rather than single results.
3. Record the right metrics per prompt

Whether your brand is mentioned, and whether each competitor is Position or prominence (first recommendation vs. buried in a list) Whether you're cited as a source (a link to your domain) versus merely named Sentiment or framing (recommended, neutral, or warned against)

4. Calculate share of voice by topic cluster For each cluster, compare your mention/citation rate to each competitor's. Clusters where a competitor's rate is well above yours are your gap topics. Sort by gap size and by business value (topics tied to purchase intent matter more than top-of-funnel curiosity).
5. Diagnose why they win For the gap topics, look at which URLs the AI cites. Common patterns: the competitor has a dedicated, well-structured page for that question; third-party sources (review sites, Reddit, Wikipedia, listicles, YouTube) favor them; or you have no content at all on the subtopic. The citation list tells you whether the fix is on your own site or in earning presence on those third-party sources. Ways to do it

Manual/DIY: a spreadsheet plus scripted API calls to the major models. Cheap, but API answers don't always match the consumer product experience, especially with web search and AI Overviews. Dedicated tools: there's a growing category of AI visibility trackers (some SEO suites have added modules, and there are standalone products). They automate prompt tracking, competitor share of voice, and citation source analysis. The landscape moves fast, so I'd check current options and pricing before committing.

Caveats Results are probabilistic and personalized, so treat them as directional trends, not exact rankings. Track over time (weekly or monthly) so you can tell real shifts from noise, and re-check after publishing content to see if the gap closes. If you tell me your industry and who your main competitors are, I can help draft a starter prompt set and a tracking sheet.