Tools that track keyword rankings in AI search results: how we report them to clients
How we report keyword rankings in AI search results for clients: turning keywords into prompts, tracking position in answer and citation rate, and what to call a rank.
Clients ask for keyword rankings in AI search because they have always had keyword rankings. It is a reasonable request and an awkward one. AI engines do not hold a fixed position for a keyword. They write an answer, and the answer changes.
Our job is to give the client something that plays the same role as a rank report without pretending it is the same thing.
Agree on the vocabulary first
We tell clients up front that "ranking" splits into three numbers.
Position in answer. If the engine lists several brands or products, where does the client sit on average? It is a mean over many runs.
Visibility and share of voice. Across a prompt set, how often is the client named compared with competitors?
Citation rate and position. How often is a client page cited as a source, and where in the list of sources?
Once the client accepts that vocabulary, reports become useful. Before that, every conversation turns into "why were we third yesterday and fifth today".
Turn keywords into prompts
Nobody types "crm startups" into ChatGPT. They type a sentence. So we convert each priority keyword into two or three natural prompts, and we tag them with the original keyword and topic so reports can roll up.
Promptwatch helps here. Prompt Explorer suggests prompts from Google Search Console, ranking keywords, People Also Ask, or any page. Prompt Volumes add search volumes, difficulty scores and citation rates across AI search, Google and Bing, which lets us prioritize prompts the same way we prioritize keywords. Prompts can be typed as organic, brand or competitor comparison, and grouped by topic and intent.
What the platform tracks
Prompt panels run daily per model, country and language. Position in Answer, Visibility Score, Share of Voice and Sentiment are all reported. We use A/B period comparison for the monthly report, so the client sees this month against last, and a heatmap compares any two dimensions, such as topics by model.
Coverage matters. Real-UI monitoring covers ChatGPT (Search), Google AI Overviews, Google AI Mode, Copilot, Gemini, Perplexity, Amazon Alexa and Meta AI. Claude, DeepSeek, Mistral and Grok are tracked through APIs, and we label that in reports. For keyword-minded clients, the Google AI Overviews and AI Mode lines get the most attention, because that is where classic search traffic overlaps.
Joining AI data with classic data
Search Insights overlays Google Search Console and Bing Webmaster data, including Microsoft Copilot, on AI crawls and citations. That gives a single view for clients who want to see classic rank next to AI visibility. For a keyword where classic rank is strong and AI citation is weak, we investigate the page. For the reverse, we look at what the engine is citing.
Citation analytics give page and domain citations with URL positions. Citation Trends show first cited, ramp-up days, peak rate and decay half-life per page. That is the best answer we have to "how long until a new page is cited".
The plumbing check
When a page that ranks well in classic search is not cited, we check Agent Analytics before anything else. IP-verified logs of GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot and others show whether the bots fetch the page and whether they hit errors. Crawler logs start on Professional.
How we write the report
One page. A table of priority keywords with their prompts, average position in answer, share of voice versus three competitors, and citation rate for the matching page. A short note on what changed and what we did. We do not publish daily numbers, because daily swings are noise.
Limits we state out loud
No tool can guarantee a position. Engines vary answers between runs, and they vary by country and language. A client who wants a fixed rank is asking for something that does not exist. We also say that volumes are estimates.
Choosing a plan
The free Explore plan covers ChatGPT only with 10 prompts. Essential is $95 a month, Professional is $245 and adds crawler logs and custom reports, and Business is $579. Agency plans start at $199 and include unlimited projects and prompts.
Other tools suit narrower needs. Suite add-ons from Semrush and Ahrefs keep AI data beside classic rank data, and AthenaHQ offers a credit-based optimization workflow. We keep the platform that joins prompts, citations, crawler logs and Search Console in one place, because that is what a keyword-minded client needs to see.