One buyer question doesn't have one fan-out. It has a whole space of them.
The library is the core deliverable clients keep. For every buyer question we work out the full set of smaller questions across nine tool profiles and three to six buyer types, then group the rows into the 30 to 60 themes that actually decide who gets quoted. We generate until we hit saturation, not a target count, because padded sets are noise sold as thoroughness.
A question library is a database of the smaller questions AI tools generate from a brand's main buyer questions. Citabld builds one across nine tools and every buyer type, labels each row with how it was obtained, and groups the rows into scored themes. The library serves as both the content plan and the measurement list.
Why nine tools instead of one list?
Because every tool breaks questions down differently, and a single generic list is wrong on eight of them. Each tool is parameterized from measured behavior rather than described in prose, so a row generated for Copilot looks nothing like a row generated for ChatGPT.
| tool | measured behavior | what it demands |
|---|---|---|
| chatgpt | rewrites heavily, changes every run, remembers you | cover phrasings the user never typed |
| perplexity | keeps your wording, very stable | classic keyword work still applies; our calibration anchor |
| copilot | compresses questions into a few keywords | keyword-dense content wins, buyer-type effects weakest |
| claude | hands back its search results mostly untouched | Brave visibility is the most direct citation lever |
| google ai mode | largest fan-outs, eight to twelve questions standard | question-level coverage, never single-run readings |
| ai overviews | similar conclusions to AI Mode, mostly different URLs | scored as a separate surface, never merged into 'Google' |
| gemini | deepest personalization, behavior-based | full buyer-type matrix |
| grok | loops, asks follow-ups from what it finds | second-order content, and X posts as a source |
| deepseek | searches nothing by default | clean baseline probe; never counted in citation metrics |
What is a buyer-type cell?
A buyer-type cell is one validated buyer profile the library generates questions for. We model six dimensions for business buyers: role and seniority, company size, geography, industry, technical sophistication, and what constrains their decision, chosen because they're the dimensions that measurably survive the AI's rewrites. Location survives almost always. Price limits often don't.
Validation before generation
Cells are validated against closed-won deals or mined reviews before they drive generation, so the matrix reflects your actual buyers rather than a marketing persona deck.
Committees and veto paths
We map the committee, because the champion's ChatGPT can be convinced while the CFO's Copilot un-convinces. Veto paths get their own cells.
One list of smaller questions is a guess. Nine tool profiles and your buyer types are a map.
Every row carries its evidence label, so you always know which questions were watched and which were modeled.
What is the Citabld Decomposition Calculus?
The Citabld Decomposition Calculus (CDC) is our named framework for working out a question's smaller-question space instead of guessing at it. The insight it's built on: the AI doesn't generate smaller questions from your question, it generates them from the answer it wants to write. CDC breaks the question into a frame, sketches what a complete answer must contain, then applies a fixed set of expansion moves to produce the smaller questions, per tool and per buyer type. Because the moves are content-free machinery and the question supplies the content, the same calculus handles any question in any industry, and it's falsifiable: every smaller question captured from a live tool either traces back to a move or grows the method.
Derived, not recalled
There's no pattern library being pasted in. Each library row is the output of a specific move applied to your question's frame, which is why the same buyer question provably yields different spaces for different buyers.
Modeled honestly
CDC also models what the AI destroys: price limits change, brand sets compress, geography survives. The library covers the post-rewrite forms, not just the pretty originals.
What does an evidence label mean?
Every row carries one of four labels describing how it was obtained. Three major tools expose no query logs to anyone, so their rows are always MODELED and always say so. This isn't a disclaimer at the bottom of a report. It's the spine of the product, because you can't make good decisions on data that hides what it is.
| label | meaning |
|---|---|
| MEASURED | captured from a live tool |
| DETERMINISTIC | follows from a documented process, like Claude's Brave setup |
| EXTRAPOLATED | derived from patents or vendor docs |
| MODELED | built from our tool profiles, checked against the ones we can watch |
What do you actually receive?
A working database, not a PDF. Every row is a question with its tool, buyer type, journey stage, evidence label, and theme. Every theme carries its scores. The same file is the content plan for production and the tracked list of prompts for measurement, which is why the library keeps paying after the engagement.
- /500 or more question rows, generated to saturation across your head queries.
- /30 to 60 theme clusters, each scored for coverage, winnability, and priority.
- /Every row labeled MEASURED, DETERMINISTIC, EXTRAPOLATED, or MODELED.
- /Journey stage tags that determine which metric is honest for each theme.
- /Yours to execute with any team, including one that is not us.
See this phase run on one of your own head queries, free.
Six questions, one modeled fan-out across seven engines, and one finished piece of content for the subquery you pick.