The questions buyers ask, answered in the first sentence.
Answer first, then the reasoning. Including the answers that cost us the sale.
Where a claim rests on a model rather than an observation, the answer says so in the same sentence.
Is query fan-out real or a theory?
Real and documented. Google describes AI Mode issuing multiple related searches per question; ChatGPT's, Perplexity's, and Copilot's retrieval queries have been independently captured and measured at scale. What varies by engine is how far the fragments drift from your question, from near-verbatim to almost complete rewrites.[src]
Can you see the exact subqueries for my brand?
On some engines, yes. We capture live fan-outs where they are observable. On Google's surfaces and Gemini, nobody can, and vendors claiming otherwise are modeling without saying so. Citabld models those engines too, and labels every such row MODELED, calibrated against the engines we can measure.
How is Citabld different from SEO?
SEO optimizes pages to rank for queries people type. Citabld engineers coverage across the fragments AI engines generate from those queries: different unit (passage, not page), different target (synthesized answers, not ranked lists), different evidence (citation share, not position). The disciplines overlap; the retrieval layer does not.
How is Citabld different from VISIBLD?
VISIBLD audits and monitors your AI visibility: where you appear, how you are framed, what changed. Citabld is the architecture and content layer that moves those numbers: the query library, the priorities, the production. They pair; each stands alone.
How long until we see citations?
Honest answer: decision-stage themes can move in weeks once coverage ships; awareness-stage influence works on the models' trained picture and moves on quarterly timescales. We tag every theme by stage and measure each with the metric that can actually move, and we show you which is which before you spend.
Do you write the content, or just tell us what to write?
We write it. Run 2 is a full content operation: long-form pillar pages, definitional blocks, comparative pages with honest verdicts, explainer and research posts, third-party placements, review programs, and structured product feeds. Each unit is drafted, sourced, edited, voice-checked, and shipped by us against the approved build sheet.
How in-depth is the content you produce?
Full depth. A priority theme becomes a complete page, structured block by block, with original information rather than restated consensus. Every unit carries a written information-gain declaration in its brief: what this passage contains that no competing passage contains. Units that cannot answer that do not get written.
How do you keep the content in our voice?
We build a Voice Card from 8 to 12 samples of your actual writing: sentence rhythm, vocabulary, and the claims you are and are not willing to make. Every unit passes a voice check against that card before it ships. Our voice never appears in client content.
What is the free fan-out, and what does it cost?
It costs nothing. Submit one head query with a short intake and Citabld generates a sample of the subquery space engines would retrieve against, clustered into themes, with every row labeled MODELED because nothing in a free sample is captured live. It is one query's slice of the Run 1 library.
Which engines can you actually measure?
We capture live fan-outs on the engines that expose them, with Perplexity serving as our calibration anchor because its fragments stay near-verbatim and stable. Google's surfaces and Gemini expose no query logs to anyone, so their rows are modeled from engine behavior profiles and labeled MODELED without exception.
Who owns the subquery library?
You do. The Query Map is delivered as a working database you can execute with any team, including one that is not us. It doubles as your tracked-prompt inventory for measurement, so it keeps reporting after the engagement ends.
Why do you score some themes as unwinnable?
Because some surfaces cannot be displaced with content. Themes owned by governments, medical bodies, or Wikipedia are entity-accuracy problems, not content problems, and we fix your facts on those surfaces instead and charge less. Naming them before you spend is the point of scoring.
Do we need VISIBLD to work with Citabld?
No. Each stands alone. VISIBLD measures where you appear in AI answers; Citabld builds the architecture and content that moves it. Running both closes the loop of audit, architecture, production, and measurement, but the Query Map and production work without it.
What do you need from us to start?
Your category, your competitors as you see them (we will correct that against the engines' view), access to buyer language such as sales calls, tickets, or the communities your buyers post in, and 8 to 12 samples of your writing for the voice standard. The mapping call scopes the rest.