Source-Grounded Content Intelligence, Simplified
A build specification for a repeatable system that produces marketing assets grounded in primary source material — reviews, transcripts, web copy, case studies, FAQs.
Derived from your Google Doc “Content Writer Details”, cataloged under Workflows & Draft Specs. Read alongside the source to trace each requirement back to your original intent.
Source-Grounded Content Intelligence, Simplified
A repeatable system that turns a source library Five Star Marketing provides and structures into high-converting, in-market assets — produced in any of eight formats (landing pages, Google Ads pages, SEO articles, blog posts, social post series, display/text ads, nurture emails, long-form guides), each shaped correctly for its format, all from one workflow with a format selector. Authentic customer voice, sequenced to convert.
Executive Finding
Your proposal is strategically right and architecturally overbuilt. Its real objective — generating high-converting marketing assets in authentic customer voice — does not require a hand-authored ontology of nodes, layers, and micro-yeses. It requires three things the platform already supplies, sequenced by one workflow.
Three components and one workflow replace seventeen sections and two phases — and that single workflow now produces any of eight in-market formats. Five Star Marketing provides and structures the source library; a Library grounds the voice, the MECLABS Methodology supplies the persuasion sequence, and a five-chapter Workflow with a format selector produces the output.
What you're actually asking for
Stripped of vocabulary, your proposal describes a single outcome: a repeatable way to produce marketing assets — landing pages, ads, SEO pages, guides, social, email — that read as your customer's own words rather than generic AI, and that are ordered to earn agreement and drive a defined action. Every one of the seventeen sections serves that one outcome.
Three intentions in your document are correct and must be preserved. We honor all three in this spec.
Authentic voice
Real customer phrasing, objections, and proof — never invented, never summarized away.
Sequenced persuasion
Copy that builds incremental agreement (micro-yeses) in a deliberate order.
Reusable system
A standardized process that scales across assets and operators, not one-off writing.
Where the design was overcomplicated
Your proposal conflates three distinct things and then tries to hand-build the two that the platform already provides. The result is a large infrastructure project standing in front of a content process that could ship now.
| In the proposal | What it really is | The overcomplication |
|---|---|---|
| Phase 1 — Raw Source Library with a 7-axis metadata taxonomy and 8 retrieval modes | A document store of full source material | Hand-tagging every source across source type, business context, persona, funnel stage, intent, geo, and performance. Semantic retrieval already does this cross-cut on demand. |
| Nodes, 8–15 persuasion layers, thousands of micro-yeses, synoptic layering, CRO overlay | MECLABS conversion methodology | Rebuilding tested conversion science as a stored object graph. The methodology is already encoded and toggleable on every Expert and App — and ships as micro-yes Apps (MY4, MY5, MY7/8). |
| Phase 2 — SOPs, agents, orchestration, dynamic assembly, contradiction detection | A guided production process | A multi-system buildout for what is, operationally, one Workflow with five steps. |
Your proposal braces, in its own words, for "dozens of node types, hundreds of persuasion layers, thousands of micro-yes patterns." That is complexity the design creates, not complexity the job requires. We decline to build it.
Three things, not seventeen
Nodes and micro-yeses are not data to store. They are generated per asset, at draft time, by methodology-grounded Experts pulling from your Library. That single move dissolves Phase 1's taxonomy and Phase 2's orchestration while keeping everything you care about.
Simple, not too simple — one workflow and five experts, but format-aware across eight outputs.
Library — grounds the voice
Your full source material, provided AND structured by Five Star Marketing — reviews, transcripts, web copy, case studies, FAQs. The system grounds every output in it and quotes it verbatim. Structuring the corpus is the client's responsibility; the system's job is to turn it into outputs.
Methodology — supplies the sequence
The MECLABS Methodology toggle injects the tested micro-yes progression at generation time. The persuasion "layers" are applied, not stored.
Workflow — sequences decisions
One Workflow Orchestrator flow with five chapters and a format selector. The operator picks the output format in Chapter 1; context carries to every downstream chapter so each expert builds to that format.
Five chapters, end to end
Each chapter has one prompt element addressed to its expert, with operator inputs collected as {{ }} form fields. The operator selects the output format in Chapter 1 from a dropdown; context carries across all five chapters. The full operational prompts are recorded in the Execution Receipt; the perspective below states each chapter's intent.
| # | Workflow Element | Perspective Prompt | Expert (built · GPT-5) |
|---|---|---|---|
| 01 | Define the Job Button + form fields. Operator selects OUTPUT FORMAT (dropdown of the eight formats) and provides audience, core problem, single action, journey stage, non-negotiables. | "You are a conversion strategist. Force clarity on the output format, the audience, the single action, the stage, and any non-negotiables. Reject vague answers. Output a one-line creative brief plus a structured brief. Do not write the asset." | Five Star Marketing - Conversion Brief Expert Methodology ON · no Library link — the brief precedes grounding |
| 02 | Mine the Voice No-input / auto. | "You are a customer-voice analyst. From the structured library only, surface verbatim phrasing, objections, proof, and emotional language — weighted to what the selected output format needs. Invent nothing." | Five Star Marketing - Voice & Proof Miner Methodology ON · Library-linked |
| 03 | Build the Persuasion Sequence No-input / auto. | "You are a MECLABS persuasion architect. Design the micro-yes sequence using the structure and conversion emphasis that matches the selected output format. Map each section to the micro-yes it earns and the proof that supports it. No final copy." | Five Star Marketing - Persuasion Architect Methodology ON · Library-linked |
| 04 | Draft the Asset No-input / auto, plus an optional length/tone field. | "You are a conversion copywriter. Build the asset in the structure the selected format requires, fusing the customer's own words and the mined proof onto the approved sequence. Flag any claim you cannot ground." | Five Star Marketing - Grounded Copywriter Methodology ON · Library-linked |
| 05 | QA & Refine Audits the draft; loops back to 04 on any fail. | "You are a skeptical QA reviewer. Audit each section on four axes — authenticity, conversion sequence, format-correctness, and brand fit. Pass/fail per section; return failures to the builder." | Five Star Marketing - Authenticity & Conversion QA Methodology ON · Library-linked |
What gets built, and in what order
The system splits cleanly into a one-time setup and a per-asset run loop. This is the entire engineering scope for v1 — deliberately small, so you see output fast and we expand only against demonstrated need.
Setup
- One Library — Five Star Marketing - Customer Voice & Proof Corpus — provided and structured by the client.
- Five Experts (Five Star Marketing - …), all on GPT-5 with Methodology ON.
- One ADS — Five Star Marketing - Content Intelligence Engine — carrying the experts and library.
- One Workflow — Five Star Marketing - Content Intelligence Workflow — of five chapters, with the output-format selector in Chapter 1.
Production loop
- Operator opens the Workflow, completes Chapter 01 (brief).
- Chapters 02–04 run with review gates between them.
- Chapter 05 QA passes or returns to 04.
- Approved asset is exported. No new infrastructure per asset.
Acceptance criteria for v1
| Criterion | Pass condition |
|---|---|
| Authentic voice | Draft uses verbatim customer phrasing from the Library; zero invented claims. |
| Sequenced persuasion | Each asset section maps to a named micro-yes in the approved sequence. |
| Format-correct | The output matches the structure the selected format requires (right sections, length, and shape). |
| Repeatability | A second operator can produce a comparable asset using only the Workflow, no tribal knowledge. |
| Time to first asset | First production-quality asset ships within the setup engagement, not after a Phase-2 buildout. |
Honored, but not in v1
Multi-format output is now IN, not deferred — the items below remain the roadmap. Your proposal's ambitions are valid as a roadmap; they are wrong as a prerequisite. The items below are explicitly deferred until v1 is in production and the data justifies them. Naming them keeps your vision intact while protecting the timeline.
Each becomes a candidate enhancement once the Library is populated, the Workflow is live, and we have asset performance to learn from. None blocks shipping.