We don't hand you a plan and wish you luck. We write the content.
Production is a full content operation, not a recommendations deck. We draft, edit, source, and ship complete pieces against the approved build sheet in two-week sprints down the priority list: long pages, short answer blocks, comparison verdicts, explainer posts, third-party placements, review programs, and structured product feeds. Every piece is written in your voice, shaped to survive being lifted, and routed to the surface where that theme's citations actually live.
Produce is the fourth phase of the Citabld method. Citabld writes the content on the approved build sheet: pages, short answer blocks, and comparisons. Each piece is drafted answer-first, checked against your Voice Card, passed through a fourteen-point quality review, and published on the surface where AI tools already cite sources for that topic.
What do you actually produce?
Finished, publishable pieces at full depth. A priority theme doesn't become a bullet in a spreadsheet; it becomes a page, and the page is written by us end to end, with the research, sourcing, and structure done before anything is styled.
- /Long pillar pages covering a theme at full depth, built block by block to be lifted.
- /Short answer blocks of 40 to 60 words, company-named and standalone, that the AI can lift without context.
- /Comparative pages with honest verdicts, including where a competitor is the better fit.
- /Explainer and research posts carrying original information rather than restated consensus.
- /Third-party placements and review programs for themes whose citations don't live on your domain.
- /Structured product feed records and trust fields for shopping fan-outs, where prose doesn't answer the query at all.
How do you keep it in our voice?
We build a Voice Card from 8 to 12 samples of your actual writing: sentence rhythm, vocabulary, the claims you'll make, and the ones you won't. Every piece passes a voice check against that card. Our voice never appears in your content, which is the whole point of collecting the samples before the first sprint.
We don't hand you a plan and wish you luck. We write the content.
Every piece passes a fourteen-point quality gate and your voice standard before it ships, and lands on the surface where the citations for that question actually live.
What is the quality gate?
A gated chain every piece moves through before it can ship. Nothing is published just because a draft looks finished.
| gate | requirement |
|---|---|
| brief | a written note on what this piece carries that no competing piece carries |
| extractable core | the correct, sourced answer drafted before any styling or narrative |
| voice pass | checked against your Voice Card, not ours |
| quality rubric | fourteen points, hard ship gate, failures return to draft |
| placement routing | published to the surface where this theme's citations actually live |
How is the content built to be quoted?
The unit the AI pulls is the passage, so we build passages. Fact-bearing blocks run roughly 40 to 60 words, in the measured sweet spot for getting pulled, the answer sits in the first sentence, the company is named inside the block, one topic per block, and claims carry dates. We apply these tactics while writing, not as a retrofit pass, because bolted-on optimization underperforms real writing.
What helps getting quoted
Statistics with linked sources, attributable quotes, and the company named inside the block all measurably raise the odds a passage is picked.
What hurts it
Keyword stuffing measurably hurts, and content hidden behind tabs and accordions gets pulled poorly, so answer blocks are never collapsed.
What is the Citabld Composition Calculus?
The Citabld Composition Calculus (CCC) is the content-side twin of the Decomposition Calculus. Where the CDC works out what the AI will ask, the CCC works out the passage that wins when it asks. The insight it's built on: content isn't written for the question, it's written for the selection. When the AI searches, it holds several candidate passages at once and picks among them, so every piece is written to win that pick: answer first, company named in-block, evidence installed, scope stated, and one axis it wins outright against the sources already being cited.
Paired to the library
Every question in the library records how it was derived, and each derivation move has a mirror writing move that answers it. The library row is the writing instruction, which is why production is mechanical rather than a matter of taste.
Pulled but not quoted has a fix
The most common failure is a passage that gets pulled and passed over. That's a writing failure with defined causes, and the CCC names them: no winning axis, diluted blocks, missing evidence, unbounded claims.
How do we know it worked?
The library doubles as the tracked list of prompts, so every shipped piece reports back with one of three outcomes, and each has a defined fix. Decision-stage themes can move in weeks. Awareness-stage influence works on the AI's trained picture and moves on a quarterly timeline, and we tell you which is which before you spend.
| outcome | fix |
|---|---|
| pulled and quoted | reinforce the theme and expand adjacent questions |
| pulled but passed over | the passage lost on substance; raise information gain and lift-ability |
| never pulled | wrong surface or wrong question; reroute placement or regenerate the theme |
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.