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 of these is a separate cognitive task. Bundling them into one prompt means each one gets a fraction of the model’s attention — and the fraction allocated to sections three through five is smaller than sections one and two, because the context window is now 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 reads the context brief and extracts six structured parameters. The voice calibration prompt reads those parameters and outputs a 200-token style specification. The structure design prompt reads the style spec and outputs a locked outline. No subsequent prompt writes freeform — every prompt executes against a tightly constrained input. The quality is consistent because the constraints are consistent.
Why 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 — degrading that prompt’s output — or you 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 in prompts 4–10. Every section prompt receives the full structure as context, which prevents sections from repeating or contradicting each other — a failure mode endemic to single-prompt generation. 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. This prevents voice drift and keeps context windows small. Third, quality pass at the end: prompt 11 reads the assembled article as a whole and flags coherence issues for prompt-level correction, not for manual editing.
Each prompt has one job. Structure before copy. Sections receive only their brief. A quality pass at the end. This is why section five reads as well as section one.
The Hub-and-Spoke Chain
The 12 prompts radiate from a central ARTICLE assembly point. Each spoke produces one deterministic artifact — a calibrated voice spec, a locked outline, a single section body — and writes it back to the central article object. The Orchestrator does not see the intermediate prompts. It sees one structured episode containing the assembled article, its metadata, and the chain’s coherence score.
Phase Breakdown
Phase 1 — Analysis (Sonnet 01–03): Brief analysis extracts six parameters. Voice calibration converts them into a 200-token style spec. Structure design produces a locked outline with section briefs. Phase 2 — Execution (Haiku 04–10): Lede with drop cap. Section bodies one through three (each receiving only its section brief plus the locked structure). Transitions. Pull quote selection. Footnote generation. Phase 3 — Quality (Sonnet 11–12): Coherence review reads the assembled article and corrects drift. Meta generation produces the description, related articles, and JSON-LD schema.
Model Routing: Sonnet vs. Haiku
The Article Library does not run all 12 prompts on the same model. It routes each prompt to the model whose capabilities match the task — Sonnet for reasoning-heavy analysis and quality review, Haiku for high-volume content execution. This is not a cost-cutting measure. It is an architectural decision that produces better output: Haiku writes section bodies more cleanly than Sonnet because its smaller, more focused attention window keeps it on task without introducing the complexity Sonnet adds when given creative latitude.
The counterintuitive finding from routing experiments: Haiku-generated section bodies score higher on voice consistency than Sonnet-generated bodies, because Sonnet’s tendency to elaborate pushes it off the locked style specification. Haiku executes the specification without editorializing. The best model for a task is not always the most capable model — it is the model whose failure modes are most compatible with the constraint structure.
Before / After: Single Prompt vs. Chain
The most direct demonstration of prompt decomposition’s value is a side-by-side comparison. Take the same article brief and run it through two pipelines: one “write the article” prompt, versus the 12-step chain. The difference is most visible at the lede, in section two (the typical degradation zone), and at the conclusion.
Single prompt
“Artificial intelligence is transforming the way businesses approach content creation. In today’s rapidly evolving digital landscape, companies are increasingly turning to AI tools to streamline their content workflows...”
12-step chain
“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.”
Single-prompt outputs degrade sharply by section two. Voice attributes that hold at 5.0/5 in the opening drop to 2.4/5 by the middle of the article. The 12-step chain maintains 4.7+ across every attribute, every section.
Voice Consistency Matrix
Voice consistency is the metric that separates AI-generated articles editors approve from ones they rewrite. The matrix scores five voice attributes across the chain — measuring how consistently each attribute holds from prompt 4 (lede) through prompt 10 (footnotes). A perfect score is 5 out of 5. The single-prompt baseline is shown for comparison.
| Voice Attribute | Lede | Body | Trans. | Close | Single |
|---|---|---|---|---|---|
| Direct register (no hedging) | 5.0 | 5.0 | 4.7 | 5.0 | 3.2 |
| Structural argument (not descriptive) | 5.0 | 4.8 | 5.0 | 5.0 | 2.4 |
| Banned vocabulary avoidance | 5.0 | 5.0 | 5.0 | 5.0 | 3.0 |
| Audience-appropriate specificity | 4.8 | 4.8 | 5.0 | 5.0 | 3.5 |
| Conclusion closes argument (not recap) | — | — | — | 5.0 | 1.4 |
Cost and Latency Breakdown
The chain produces a complete article in 90–110 seconds at approximately $0.017 per run. Sonnet’s three analysis prompts take ~25 seconds combined. Haiku’s seven execution prompts take ~55 seconds. Sonnet’s two quality prompts take ~20 seconds. At 100 articles per month, the routing optimization saves approximately $3.10 versus running everything on Sonnet — modest in isolation, but compounding across multiple libraries in the IO Platform.
The cost gap matters less than the quality gap. A single Sonnet prompt for the same article scores 3.2/5 on voice consistency. The hybrid chain scores 4.9. The chain does not make the models smarter. It gives each step the context budget to do its job well.