How can I track competitor visibility in AI search? The method we use for clients
How to track competitor visibility in AI search: the prompt set, share of voice, citation and sentiment method we use for clients, and the platform features behind it.
Every client wants to know the same thing within the first week: who is beating us in AI answers? It is a good question and an easy one to answer badly. A screenshot of one ChatGPT reply proves nothing. Here is the method we use to track competitor visibility in AI search, and how we run it.
Start with who actually competes
Ask the client for their competitor list. Then check it against what the engines say. In nearly every program we run, an unexpected brand shows up: a marketplace, a media site, a niche tool. Those are competitors for the purposes of AI answers, even if they are not competitors in sales.
Promptwatch helps because competitors are auto-discovered from the responses, and brand merging combines the aliases so one company is not counted three times. Brand aliases and a Brand Book, whose analysis instructions decide what counts as a mention, keep the measurement consistent over time. We set those up in week one, because changing them later breaks trend lines.
Build a prompt set that mirrors buyers
We write 30 to 50 prompts per client across four types: category, comparison, problem and brand. Each gets a topic tag and, where it matters, a persona. Persona stacking lets us ask the same question as different buyers. Prompt Explorer can suggest prompts from Search Console, ranking keywords, People Also Ask or a page, so we are not starting from our own assumptions.
Measure four things
We report four numbers per competitor, per topic.
- Mention rate. How often the brand is named across runs.
- Share of voice. Mentions as a fraction of all brand mentions in the set.
- Position in answer. Where the brand sits when several are listed.
- Sentiment. How the engine describes it, scored per response with a weekly summary.
Prompt panels run daily per model, country and language. That matters, because a competitor can dominate in one country and be absent in another. Real-interface monitoring covers ChatGPT (Search), Google AI Overviews, Google AI Mode, Copilot, Gemini, Perplexity, Amazon Alexa and Meta AI. Claude, DeepSeek, Mistral and Grok come through APIs, and we label them.
Use the heatmap, not a spreadsheet
The competitor heatmap compares any two dimensions: brands by model, brands by topic, topics by country. It is how we find that a competitor owns one topic on one engine while we are even elsewhere. That is far more actionable than an overall share-of-voice number. A/B period comparison then shows movement month to month.
Explain the gap with sources
The numbers say who wins. Sources say why.
Citation analytics list the pages and domains engines cite, with URL positions. Page Tracker follows specific URLs, including competitor pages, so we can see when a rival's guide is cited and ours is not. Reddit, YouTube and LinkedIn citation reports isolate community and video sources. Offsite mentions show where a competitor is named on third-party pages.
Paid placements matter too. Ads Radar captures sponsored ads inside ChatGPT Search, Google AI Overviews and AI Mode on tracked prompts, showing advertiser, copy and landing page. It starts on Business and is on every agency plan. It does not cover Perplexity.
Check the plumbing before blaming content
If a competitor is cited and the client is not, we check whether AI crawlers reach the client's pages. Agent Analytics gives IP-verified logs for GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot and others. Crawler logs start on Professional.
What we do not claim
We do not claim to control the outcome. Answers vary, and competitor numbers swing. We report four to eight weeks of trend, and we tell clients that a competitor's lead can shrink or grow for reasons neither of us sees.
Reporting
A one-page summary: share of voice for the client and three named competitors, biggest movers, top three citation gaps, and one action list. Scheduled PDF reports and Agent Chat help produce it, and the Action Board holds the follow-up tasks.
Where other tools fit
AthenaHQ and Scrunch AI both offer competitor tracking, and suite add-ons from Semrush and Ahrefs keep it beside classic SEO data. Heavily sampled enterprise platforms exist for executive benchmarks. We keep returning to the platform that joins the heatmap, sources, crawler logs and ads in one place.
Start free on ChatGPT only with 10 prompts, then move to Essential at $95 a month, Professional at $245, or an agency plan from $199.