Library Deep DiveIO Content Ops Series · Article 03

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

12 prompts. Drop cap to footnote. How the IO Article Library chains decomposed prompts to produce publication-ready editorial without a single revision loop — and exactly which model handles each step.

Tommy Saunders
Founder, IntelligentOperations.ai
March 29, 202610 min read
IO-CB-2026-001 · SERIES PLAN · A03 · MARCH 2026
Direct Answer
How does the IO Article Library produce publication-ready long-form content?
The IO Article Library runs 12 sequential prompts, each scoped to one task: brief analysis, voice calibration, structure design, lede, three section bodies, transitions, pull quote, footnotes, a quality pass, and meta generation. Prompts 1–3 run on Claude Sonnet for reasoning-heavy analysis. Prompts 4–10 run on Claude Haiku for fast execution. Prompts 11–12 return to Sonnet for coherence review. The full chain completes in 90–110 seconds and is returned to the Orchestrator as a single structured episode — no revision loops, no follow-up prompts.
JSON-LD SchemaSource: IntelligentOperations.ai · IO Platform · March 2026

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 6 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. This means 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.

Tommy Saunders · Founder, IntelligentOperations.ai

The 12-Prompt Chain — Interactive

The indigo nodes run on Claude Sonnet 4 (reasoning-heavy). The green nodes run on Claude Haiku (fast execution).

Article Library — 12-Prompt Sequential Chain
Sonnet
Haiku
Phase 1 — Analysis (Sonnet 4)
Phase 2 — Execution (Haiku)
Phase 3 — Quality Pass (Sonnet 4)
OUTPUT
Episode JSON
~48 token delta

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 faster and 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.

Model Routing Architecture — Article Library
Claude Sonnet 4
Reasoning · Analysis · Quality
4 Prompts
P01Brief Analysis — extract 6 params
P02Voice Calibration — derive style spec
P03Structure Design — locked outline
P11Coherence Review — full quality pass
Claude Haiku
Execution · Speed · Volume
8 Prompts
P04Lede + Drop Cap paragraph
P05Section 1 body copy
P06–07Section 2 & 3 body copy
P08Transitions between sections
P09–10Pull quote + Footnotes
P12Meta description + Related
Cost Comparison — All-Sonnet vs. Hybrid Routing
All Sonnet
$0.048
Hybrid (IO)
$0.017
Cost reduction via hybrid routing:~65%

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. Below are outputs for the same article brief — one generated with a single “write a full article” prompt, one generated through the 12-step chain.

Output Comparison — Lede
✗ 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 and achieve greater efficiency.
▼ Generic opener. “Today's rapidly evolving digital landscape” is a tell. Leads with AI tools, not the reader's problem.
✓ 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.
▲ Opens with specific professional recognition. Makes a structural claim. Uses the voice spec's direct register.

Voice Consistency Matrix

Voice consistency is the most visible quality signal in long-form content. The 12-prompt chain maintains voice consistency because each section prompt receives the style specification (output of prompt 2) as its primary input — not the growing conversation. The result: measurably consistent voice from the lede to the conclusion.

Social Distribution Suite

Tommy Saunders
@tommysaunders_io
Why does AI writing quality degrade after section two? Because a single prompt is doing 12 jobs at once. The IO Article Library splits those into 12 separate prompts: • 3 on Sonnet (reasoning) • 7 on Haiku (execution) • 2 on Sonnet (quality pass) Result: section 5 reads as well as section 1.
38 replies201 reposts467 likes

Search Package — SEO + AEO

SEOSearch Package Preview
intelligentoperations.ai › content-ops › article-library
Inside the Article Library: How the Writing Engine Produces Long-Form at Scale
12 prompts. Drop cap to footnote. How the IO Article Library chains decomposed prompts to produce publication-ready editorial without a single revision loop.
Answer Engine Optimization
How does prompt decomposition improve AI writing quality?
Prompt decomposition breaks a complex writing task into 12 single-responsibility prompts. Structure is locked before body copy is written. Each section receives only its brief, not prior sections. A quality pass at the end checks coherence. The result is consistent quality from paragraph one to the conclusion.
ai article writing systemprompt chain contentlong-form ai contentarticle library workflowprompt decomposition
CRM Integration

Get the 12-Prompt Chain Template

Download the complete prompt chain architecture including model routing, token budgets, and prompt templates for building your own decomposed writing system.

No spam. Unsubscribe anytime. We respect your privacy.
Nurture Sequence
Day 0
Welcome + Chain Architecture PDF
Day 3
Prompt Decomposition Guide
Day 7
Model Routing Comparison
Day 14
Before/After Case Study
Day 21
Build Your Own Chain Session

Frequently Asked Questions

5 questions
The IO Article Library runs 12 sequential prompts, each scoped to a single task: brief analysis, structure design, lede writing, section bodies (one per section), pull quote selection, footnote generation, meta description, and related article suggestions. Each prompt receives only its necessary inputs — not the entire prior conversation — which keeps quality consistent from the first section to the last. The full chain runs in under two minutes on Haiku with Sonnet quality passes at the beginning and end.
Prompt decomposition means breaking a complex task — like writing a 2,500-word article — into discrete, single-responsibility prompts that each produce one deterministic output. Instead of ‘write me a good article about X,’ a decomposed chain runs: analyze the brief, design the structure, write the lede, write section one, write section two, select the pull quote, generate the meta description. Each prompt is easier for the model, produces a better output, and fails gracefully — a weak section two doesn’t contaminate section three.
The IO Article Library uses a two-model routing strategy. Prompts 1 through 3 (brief analysis, voice calibration, structure design) run on Claude Sonnet 4 — these are reasoning-heavy tasks where quality matters most. Prompts 4 through 10 (section bodies, transitions) run on Claude Haiku — faster and cheaper for high-volume execution tasks. Prompts 11 and 12 (quality pass, meta generation) return to Sonnet for a final coherence review. This routing reduces cost by approximately 65% versus running all prompts on Sonnet.
A single ‘write me a 2,500-word article’ prompt produces inconsistent quality: strong openings that weaken by section three, voice drift across sections, and no structural control. The 12-prompt chain solves each of these: structure is locked before any body copy is written, each section prompt receives only the section brief plus the locked structure (not prior sections), and a final quality pass checks coherence. The result is quality that stays consistent from paragraph one to the conclusion — which is impossible with a single prompt.
The Article Library completes a full 12-prompt chain — producing a publication-ready article of approximately 2,000 to 2,800 words — in 90 to 110 seconds. Prompts 1 through 3 on Sonnet take approximately 25 seconds combined. The 7 execution prompts on Haiku run in approximately 55 seconds. The 2 quality-pass prompts on Sonnet take approximately 20 seconds. The full article is assembled and returned to the Orchestrator as a single structured episode in under two minutes.
References
1IO Article Library Architecture Documentation, v2.1. Internal technical specification for the 12-prompt sequential chain and model routing strategy.
2Saunders, T. (2026). "Prompt Decomposition for Long-Form AI Writing." IO Technical Architecture Series.
Tommy Saunders
Founder, IntelligentOperations.ai
Building AI-native operations for commercial real estate. Writing about the systems that build the systems.
Article Series