See the query fan-out behind every AI answer
Query Fan-Out shows the searches an engine runs before it answers. Those smaller searches decide what the answer says. Citlyze captures them for your tracked prompts, so you optimise for the queries that pick the winners.
14-day free trial · No credit card required

Trusted by marketing, SEO and agency teams.
- ChatGPT
- Gemini
- Perplexity
- DeepSeek
- Grok
- Google AI Overview & AI Mode
- Bing Copilot

The real queries providers ran
For every tracked prompt, see the sub-queries each engine reported while building its answer, grouped by prompt and provider.
- Fan-out queries captured per prompt and engine
- Brands named inside the queries, yours and competitors'
- Included in every plan, no setup needed
Tracked prompt
What is the best CRM for a small team?
Coverage where providers stay silent
Some engines expose nothing about their searches. Citlyze models likely variants for those prompts and labels them as modeled, kept apart from observed data.
- Modeled variants for engines that report nothing
- Comparison, follow-up, and reformulation query types
- Labeled modeled and kept apart from observed queries
SERP gap
Google rankings for fan-out queries · checked weekly
Where Google ranks you on those queries
Fan-out queries are real Google searches. A weekly rank check shows who owns them today, and where you are absent.
- Weekly Google rank check for top fan-out queries
- See which fan-out queries competitors rank for
- 25 or 50 tracked queries, depending on plan
Query fan-out is what AI engines do before answering: they break a prompt into several smaller searches, run them, and build the answer from those results. Those queries shape the answer, so tracking them shows what shaped it.
A prompt like "what is the best CRM for a small team" typically fans out into searches such as a comparison query, a pricing query, and a review query for specific brands. Citlyze shows the actual set per tracked prompt and engine, so you work from real examples rather than guesses.
Ranking for the visible prompt is no longer the whole game: the pages an engine reads come from the fan-out queries. Content that ranks for those sub-queries gets read, cited, and paraphrased into answers. The SERP gap view exists exactly for this: it shows which fan-out queries you rank for on Google and which competitors own.
The same mechanism applied by LLM-based assistants: ChatGPT, Gemini, and others expand a prompt into search queries before composing a response. Different engines fan out differently and report it with very different fidelity, which the coverage view makes explicit.
Fan-out queries are a ready-made cluster plan. The recurring themes across your prompts' fan-out show which subtopics engines keep searching for, and a page per recurring sub-query is a cluster built from observed demand instead of keyword-tool guesses.
Captured queries are the ones a provider reported while answering your tracked prompt. Modeled variants are Citlyze's predictions for engines that expose nothing. The two are labeled and kept apart, so you know whether you are looking at an observation or an estimate.
Yes. Google has described AI Mode as issuing multiple related searches behind a single question, and its answers draw on those results. Google does not expose that fan-out per answer, which is where modeled variants and the weekly Google rank check fill the gap.
Add your buyer prompts to a Citlyze workspace and fan-out fills in automatically as tracking runs: no extra setup. Captured queries appear per prompt and engine, and the insights view aggregates themes, brand mentions inside queries, and the SERP gap.
The weekly Google rank check covers your most frequent fan-out queries: 25 on Growth and 50 on Pro and above. Agency plans cover 25 per workspace. Captured and modeled fan-out itself is included in every plan.
The free simulator models what a prompt would likely fan out into; it is a simulation and says so. The product captures what engines ran for your tracked prompts, adds modeled coverage where providers report nothing, and checks Google rankings for those queries weekly.
Does AI recommend you when buyers ask?
Track how ChatGPT, Gemini, Perplexity, DeepSeek, Grok, Google AI Overview & AI Mode, and Bing Copilot mention, cite, and rank your brand. Review tracked answers and the gaps to close.
No credit card required
