Artifact types
WSD (Workflow State Document)
Knowledge Base (voice, argument bank, pub intel)
Pipeline admin (registry, calendar, tracker)
Source management (Zotero, notes, confidence)
Published / distributed output
Actions
CREATE New artifact produced
READ Consumed as input
UPDATE Existing artifact modified
ARCHIVE Retired / filed
Phase 1
Strategize
Decide what to write, for whom, and why it matters to the business
Inputs
READ Editorial Calendar
READ Project Registry
READ Publication intelligence profiles
MISSING Content Strategy document
MISSING Sector Intelligence database
MISSING Content performance data (from Phase 7)
→
Outputs
UPDATE Editorial Calendar (priority decisions)
MISSING Content Strategy doc (updated quarterly)
MISSING Commissioned Brief (triggers Phase 2)
Phase 1 produces no formal artifact. Strategic decisions live only in Richard's head or informally in the calendar. No document triggers Phase 2 -- it begins when Richard says so, not when a file signals readiness. Rationale for piece selection is never recorded.
Phase 2
Commission
Define the brief, choose the pattern, initialize the project
Inputs
READ WSD Template
READ Pattern library (PATTERN-*-v1)
READ Voice Profile (register selection)
READ Publisher Disclosure Matrix
READ Project Registry (ID assignment)
→
Outputs
CREATE WSD -- S1 Locked Commitments (stub)
CREATE WSD -- S7 Project Metadata
CREATE Project folder (/projects/[PROJECT-ID]/)
UPDATE Project Registry (new entry)
UPDATE Editorial Calendar (new entry)
Phase 3
Produce
Research, draft, verify, revise -- the core writing work
Stage 1
Topic & Angle
READ WSD S1 (stub)
READ Voice Profile B6 (audience calibration)
READ Context Guide (practitioner / linkedin)
READ Editorial Calendar (topic)
→
UPDATE WSD S1 -- thesis, audience, register, constraints locked
UPDATE WSD S1 -- Discourse Scan
UPDATE WSD S1 -- Experience Intent, Conclusion Intent, Evidence Selection
Stage 2
Outline
READ WSD S1 (Locked Commitments)
READ Argument Bank
READ Theoretical Framework Working Notes
→
UPDATE WSD S1 -- Approved Outline (locked)
UPDATE WSD S2 -- Current Stage
Stage 3
Draft
READ WSD S1 (Locked Commitments + Outline)
READ Voice Profile Section P (full)
READ Context Guide (channel-specific calibration)
READ Source Registry + Reading Notes
READ Discourse Ground (ProActive / personal brand)
→
CREATE [TITLE]_AI_draft_v1.md
UPDATE WSD S5 -- Output Artifacts (entry)
UPDATE WSD S8 -- Session Brief
Stage 4
Review & Revision
READ AI Draft v1 (Richard reviews)
READ Voice Profile (authenticity check)
READ WSD S1 -- Locked Commitments (constraint check)
→
CREATE [TITLE]_v2.md (human-reviewed)
UPDATE WSD S4 -- Completed Stages
UPDATE WSD S6 -- Decision Log
Stage 5
QA & Finalize
READ Revised draft v2+
READ Confidence Schema
READ Verification Pipeline
READ Pub Intelligence profile (venue requirements)
→
CREATE ARGUMENTATION_ANALYSIS.md
CREATE REGISTER_CHECK.md
CREATE Final draft (publication-ready)
UPDATE WSD S5 -- Output Artifacts (final)
MISSING Visual Brief (to trigger Phase 4)
Stage 5.5
Atomization
READ Final draft
READ WSD S5b -- Distribution Plan
READ Context Guide -- LinkedIn (spoke calibration)
→
CREATE LinkedIn post(s) -- atomization spokes
CREATE Email excerpt draft
UPDATE WSD S5b -- Distribution Plan (spokes listed)
MISSING Distribution Checklist (spoke tracking)
Stage 5 produces no Visual Brief -- a spec naming required visuals, data to be visualized, and brand context. Phase 4 has no governed input artifact and so never gets triggered. The Distribution Plan (WSD S5b) captures channels but no artifact instantiates and tracks spoke execution.
Phase 4
Visualize
Create visual assets for the content
Inputs
READ Final article text
MISSING Visual Brief (from Stage 5)
MISSING Brand Asset Library (governed)
→
Outputs
MISSING Hero image
MISSING Social media cards (per channel)
MISSING Infographic (for data-driven pieces)
Phase 4 is entirely ungoverned. Tools exist (Canva, Pencil, AI image gen) but no workflow, no input artifact, and no output tracking. Visual production is ad hoc.
Phase 5
Publish
Format and post to primary channel
Inputs
READ Final draft
READ WSD S7 -- Project Metadata
READ Publisher Disclosure Matrix
READ Pub Intelligence profile (format spec)
→
Outputs
CREATE DISCLOSURE_STATEMENT.md
CREATE Formatted submission package
CREATE Published artifact (live URL / post)
UPDATE Submission Tracker
UPDATE WSD S8 -- Session Brief (final)
MISSING Published-status signal (to trigger Phase 6)
Publication produces no artifact that signals "begin distribution." The Distribution Plan in WSD S5b must be manually consulted -- it is not triggered by publication status. The handoff to Phase 6 depends entirely on Richard remembering to initiate it.
Phase 6
Distribute
Push to secondary channels, repurpose, amplify
Inputs
READ Published article (live URL)
READ WSD S5b -- Distribution Plan
READ Context Guide -- LinkedIn
READ Atomization spokes (from Stage 5.5)
MISSING Distribution Checklist (per content type)
→
Outputs
CREATE LinkedIn posts (published)
CREATE Email newsletter excerpt (published)
MISSING Proposal language bank entry
MISSING Distribution completion record
Phase 7
Measure
Track performance and business impact
Inputs
MISSING LinkedIn analytics data
MISSING Website analytics (GA4)
MISSING Email platform metrics
→
Outputs
MISSING Monthly performance report
MISSING High/low performer flags
MISSING Performance signal for Phase 1 (closes loop)
Phase 8
Learn & Loop
Feed learnings back into strategy and production
Inputs
READ Completed WSD (full project record)
MISSING Performance report (from Phase 7)
→
Outputs
UPDATE Failure Library
UPDATE Pattern Retrospectives
UPDATE Outcome Analysis
UPDATE Voice Profile calibration log (if signals found)
ARCHIVE WSD (project complete)
MISSING Content Strategy update (loop back to Phase 1)
What This Map Reveals
Gap 1 -- Phase 1 produces no artifact
Strategic decisions never become a document. There is no Content Strategy or Commissioned Brief that formally initiates a project or records why this piece was selected. Phase 2 begins when Richard says so -- not when a file signals readiness. The rationale for piece selection is permanently unrecorded.
Gap 2 -- Stage 5 produces no Visual Brief
Phase 3 (Produce) ends with a final draft but produces no artifact that triggers Phase 4 (Visualize). A Visual Brief -- naming required visuals, data to illustrate, and brand context -- needs to be a Stage 5 output. Without it, Phase 4 has no governed input, and visual production is permanently ad hoc.
Gap 3 -- Publication triggers nothing
When a piece is published, no artifact in the system changes state. The WSD Distribution Plan exists but must be manually consulted. There is no published-status signal, no Distribution Checklist that instantiates per piece, and no completion record that closes the distribution loop. Phases 5 to 6 is the most broken handoff in the studio.
Strong -- WSD is the model for every handoff
The Phase 2 to 3 handoff works because the WSD is a continuously updated state document that carries complete context across agents and sessions. Locked Commitments (S1) + Session Brief (S8) give any cold-start agent everything it needs. The three broken handoffs above all need their own WSD-equivalent: a living artifact that captures state, triggers the next phase, and is updated on completion.
Insight -- The knowledge base is the hidden backbone
The Knowledge Base is read at Stage 1 (audience calibration), Stage 2 (argument structure), Stage 3 (voice + source), Stage 4 (authenticity check), and Stage 5 (venue requirements). It is consumed at nearly every sub-stage but only written back to in Phase 8. Voice Profile and Argument Bank are the highest-leverage inputs in the system -- the quality of every draft depends on their currency.
Partial -- Phase 8 learns from process but not performance
The intelligence layer (Failure Library, Pattern Retrospectives, Outcome Analysis) actively captures process learning at Phase 8. But without a performance report from Phase 7, it cannot capture outcome learning -- which pieces resonated, which sectors respond, which arguments land. The loop is half-closed: process improves, strategy does not.
Artifact Flow Map v1.0 -- March 2026 -- Writing Production Studio