PlatformQuery Fan-Out
see how
AI models
actually search.
Behind every answer is a chain of micro-searches the model runs against the open web. Centium captures every branch so you can see what it’s looking at, not just what it ends up saying.
one prompt. many searches.
When AI models receive a prompt, they break it down into micro-searches.
Centium reports the entire chain of searches from ChatGPT and Gemini, giving you insight into the exact search process behind every answer. It’s the data foundation for AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization).
Every branch is an opportunity. To rank, to be cited, to influence the answer. Or to be invisible.
seo measures one query.
ai answers from many.
Your seo team is optimizing the tip of an iceberg.
Centium reveals how AI models actually probe a topic, just like a human would. Reviews, rankings, news, competitor comparisons, fit and use-case questions, all the angles the model considers before it answers.
Each branch is a content opportunity, an SEO, AEO, and GEO opportunity, a chance to be the source the model cites.
what we capture.
the original prompt.
Centium helps generate prompts that matter to your business, grouped within your key product segments. We push these to the major AI models to trigger their fan-out process.
every micro-search.
Every sub-query the model ran to form its answer, captured per model so you can see exactly how ChatGPT and Gemini differ on the same prompt.
brand & sources.
Centium aggregates every micro-search behind a prompt and reports whether your brand surfaced in the answer, plus the sources the model referenced along the way.
fan-out a prompt.
built for content teams.
The full fan-out, exported as a flat sheet your content team can brief from directly. Drop it straight into your preferred LLM to draft pages, briefs, or category-page copy that matches how AI actually searches.
- One row per micro-search. Category, prompt, model, and the exact sub-query the model issued.
- CSV and XLSX. Easy access to the full fan-out in either format, ready to drop into Claude, ChatGPT, or whichever LLM you use to build website content.
- Re-run on cadence. Every refresh stamps a snapshot, so you can diff fan-outs week over week.
| Category | Prompt | Model | Sub-Query |
|---|---|---|---|
| Trail & Outdoor | What are the best trail running shoes for women in 2026? | ChatGPT | top-rated trail running shoes 2026 women |
| Trail & Outdoor | What are the best trail running shoes for women in 2026? | ChatGPT | Demo Athletic Trailrunner Pro specs |
| Trail & Outdoor | What are the best trail running shoes for women in 2026? | ChatGPT | best women's trail shoes for technical terrain |
| Trail & Outdoor | What are the best trail running shoes for women in 2026? | ChatGPT | Hoka Speedgoat 6 women review |
| Trail & Outdoor | What are the best trail running shoes for women in 2026? | ChatGPT | trail shoes for narrow feet women |
| Trail & Outdoor | What are the best trail running shoes for women in 2026? | ChatGPT | cushioned vs minimalist trail shoes |
| Trail & Outdoor | What are the best trail running shoes for women in 2026? | Gemini | top women trail shoes 2026 buyer guide |
| Trail & Outdoor | What are the best trail running shoes for women in 2026? | Gemini | Hoka vs Demo Athletic women trail |
| Running | I just signed up for my first marathon. What gear do I need? | ChatGPT | beginner marathon training plan gear list |
| Running | I just signed up for my first marathon. What gear do I need? | Gemini | best running shoes for first marathon |
how the capture works.
send a prompt.
You pick the prompts that matter, or use ours: industry-tested questions per cycle, mapped to your categories.
mirror the fan-out.
We run each prompt against ChatGPT and Gemini and extract the fan-out: every sub-query the model issues on its way to the answer.
log the chain.
Each micro-search is logged with its parent query, category, model, and the order the search ran in.
hand to content.
The full chain exports as a flat sheet your team can drop into an LLM to generate briefs, outlines, and content that maps to every angle AI looks for.
same prompt.
two trees.
Different models fan out differently. ChatGPT goes wide; Gemini goes deep. Both miss things the other catches. We capture both, so you can optimize for every angle AI may search for.
ChatGPTGeminisee every search
behind your brand.
The fan-out for your category, captured on a cadence, exportable to your content team. Skip the keyword-tool guess and brief from the chain your customers actually set off when they ask AI.