Inside the Article Library: How the Writing Engine Produces Long-Form at Scale

Every senior editor knows the feeling: you ask someone to write a 2,000-word article and get back something that opens brilliantly, coasts through the middle, and ends on a sentence that sounds like the writer ran out of energy and caffeine at the same moment. That is not a people problem. It is an architecture problem.

A single “write me a great article about X” prompt hands the model too many responsibilities at once: understand the brief, choose a structure, establish a voice, write a compelling lede, maintain quality across 2,000 words, end well. Each is a separate cognitive task. Bundling them means each gets a fraction of the model’s attention — and the fraction allocated to sections three through five is smaller, because the context window is full of everything that came before.

The IO Article Library solves this with prompt decomposition. Each of the 12 prompts has one job. The brief analysis prompt extracts 6 structured parameters. The voice calibration prompt outputs a 200-token style specification. The structure design prompt outputs a locked outline. No subsequent prompt writes freeform.¹

01Why 12 Prompts and Not One

The number 12 is not arbitrary. It is the result of decomposing a publication-ready article into its minimum set of non-overlapping, single-responsibility tasks. Remove any one prompt and you either push its work onto an adjacent prompt or skip the step entirely — producing a detectably worse article.

The key decomposition decisions are three. First, structure before copy: the outline is locked in prompt 3 before any body copy is written. Second, sections receive only their brief: each section-body prompt receives the locked outline and its specific section brief, not the full text of prior sections. Third, quality pass at the end: prompt 11 reads the assembled article as a whole and flags coherence issues for prompt-level correction.²

02The 12-Prompt Chain — Interactive

Click any step to see its model assignment, token budget, and role. The indigo nodes run on Claude Sonnet 4 (reasoning-heavy). The green nodes run on Claude Haiku (fast execution).

03Model Routing: Sonnet vs. Haiku

The Article Library routes each prompt to the model whose capabilities match the task — Sonnet for reasoning-heavy analysis and quality review, Haiku for high-volume execution. This is not cost-cutting; it produces better output. Haiku writes section bodies more cleanly because its focused attention window keeps it on the locked style spec.

The counterintuitive finding: Haiku-generated section bodies score higheron voice consistency than Sonnet, because Sonnet’s tendency to elaborate pushes it off the locked spec. The best model for a task is not always the most capable — it is the one whose failure modes are most compatible with the constraint structure.

04Before / After: Single Prompt vs. Chain

The same article brief — one generated with a single “write a full article” prompt, one through the 12-step chain. Tabs show the Lede, Section 2 (the degradation zone), and the Conclusion.

05Voice Consistency Matrix

The matrix scores five voice attributes across the chain — measuring how consistently each holds from prompt 4 (lede) through prompt 10 (footnotes). A perfect score is 5.0. The single-prompt baseline is shown for comparison.

The baseline collapses most dramatically on “conclusion closes argument” — 1.4/5 — because a single prompt, having written 1,800 words, has almost no context window left for strategic thinking. The chain dedicates an entire Sonnet prompt (P11) to reading the complete article and writing a conclusion that responds to its own lede.

06Social Distribution Suite

07Search Package — SEO + AEO

08CRM Lead Capture + Nurture