Phase 1
Strategize
Decide what to write, for whom, and why it matters to the business
Richard
- Set content priorities based on business goals
- Identify which sectors need attention
- Choose between intellectual interest vs. business need
- Approve content calendar
DECISION MAKER
Current: ~5% of time | Target: ~20%
AI Agents (TARGET)
- Scan sector trends and competitor publications
- Surface content gaps and opportunities
- Analyze past performance to recommend topics
- Draft quarterly content calendar proposal
NOT YET BUILT
Systems (TARGET)
- Content Strategy document (living)
- Audience/Stakeholder Map
- Sector Intelligence database
- CRM data on client conversations
NOT YET BUILT
Phase 2
Commission
Define the brief, choose the pattern, initialize the project
Richard
- Define topic, angle, core argument direction
- Choose target venue / channel
- Set HITL mode for project
- Approve WSD Locked Commitments
AUTHOR / DIRECTOR
Current: ~10% | Target: ~10% (appropriate)
AI Agents
- Create project folder from template
- Initialize WSD with pattern
- Load publication intelligence profile
- Generate initial research brief
OPERATIONAL
Systems
- Project Registry auto-updated
- Editorial Calendar entry created
- WSD template instantiated
- Pattern library queried
OPERATIONAL
Phase 3
Produce
Research, draft, refine, verify — the core writing work
Richard
- Stage-gate approvals at each boundary
- Intellectual authority on arguments
- Voice calibration and authenticity check
- Bias Mitigation Protocol participation
AUTHORITY / APPROVER
Current: ~60% of time | Target: ~30%
AI Agents
- Literature search (3-phase protocol)
- Drafting (AI-driven or Collaborative mode)
- Citation verification (/verify-citations)
- Coherence review, gap audit
- Source confidence scoring
OPERATIONAL — STRONG
Systems
- Zotero: reference management
- Research Rabbit + Elicit + Scite
- Knowledge Base: voice, arguments
- WSD: state tracking
- Flag system: [VERIFY], [MISATTRIBUTION], etc.
OPERATIONAL — STRONG
Phase 4
Visualize
Create accompanying visual assets for the content
Richard
- Approve visual direction
- Brand compliance final check
APPROVER
Current: ~5% | Target: ~5% (appropriate)
AI Agents (TARGET)
- Generate hero image options
- Create social media cards
- Build infographic from article data
- Extract framework into slide deck
- Apply brand guidelines automatically
PARTIALLY AVAILABLE — NO WORKFLOW
Systems
- Canva: templates and brand kit
- Brand Asset Library: logos, colors
- Pencil: design generation
- AI image gen: constrained by brand rules
TOOLS EXIST — UNGOVERNED
Phase 5
Publish
Format and post to primary channel
Richard
- Final approval to publish
- Publisher disclosure sign-off
- Manual posting to most channels
MANUAL — BOTTLENECK
Current: ~10% | Target: ~5% (approval only)
AI Agents
- Format for target venue
- Generate AI disclosure statement
- SEO metadata (blog posts)
PARTIAL
Systems (TARGET)
- CMS with publish workflow
- Automated formatting per channel
- Content indexed and tagged on publish
- Archive entry auto-created
MOSTLY MANUAL
Phase 6
Distribute
Push to secondary channels, repurpose, amplify
Richard
- Personal LinkedIn engagement
- Share in relevant conversations
- Forward to specific contacts
AD HOC
Current: ~5% | Target: ~10% (strategic engagement)
AI Agents (TARGET)
- Execute repurposing cascade (8-12 derivatives)
- Adapt voice per channel
- Schedule posts across platforms
- Draft email newsletter excerpt
- Create proposal language bank entry
NOT YET BUILT
Systems (TARGET)
- Distribution checklist (per content type)
- Social media scheduler
- Email platform integration
- Cross-channel tracking
NOT YET BUILT
Phase 7
Measure
Track performance and business impact
Richard (TARGET)
- Review monthly performance dashboard
- Note which content clients reference
- Connect content to business outcomes
NOT YET BUILT
Current: ~0% | Target: ~10%
AI Agents (TARGET)
- Collect metrics across all channels
- Generate monthly content performance report
- Flag high and low performers
- Correlate content topics with engagement
NOT YET BUILT
Systems (TARGET)
- LinkedIn analytics API
- Website analytics (GA4 or equivalent)
- Email platform analytics
- CRM inquiry source tracking
NOT YET BUILT
Phase 8
Learn & Loop
Feed learnings back into strategy and production
Richard (TARGET)
- Adjust content strategy based on data
- Reallocate effort to high-performing sectors
- Evolve thought leadership positioning
NOT YET BUILT
Current: ~0% | Target: ~10%
AI Agents
- Update Failure Library (process)
- Pattern Retrospective
- Signal Log entry
PROCESS LEARNING ONLY
Systems
- Intelligence Layer (exists)
- Continuous Excellence Protocol
- Component Registry
EXISTS — PROCESS ONLY
The Time Rebalancing Opportunity
Richard currently spends ~60% of content time in Phase 3 (Production) — drafting, reviewing, refining. The Writing Studio's AI infrastructure should reduce this to ~30%, freeing 30 percentage points for strategic work (Phase 1) and performance analysis (Phases 7-8). This isn't about working less; it's about working on the activities that only Richard can do: setting direction, exercising intellectual judgment, and interpreting business impact.
The Handoff Black Holes
Three handoffs currently have no defined process: (1) Phase 3 -> Phase 4 (article done, now what visuals?), (2) Phase 5 -> Phase 6 (published, now distribute where?), (3) Phase 7 -> Phase 1 (learned something, now change what?). Each of these is a point where momentum dies and content value is lost.
Where AI Is Underutilized
AI agents are heavily used in Phase 3 (production) and barely exist in Phases 1, 4, 6, and 7. The irony: the phases where AI could have the most business impact (trend scanning, repurposing, performance analysis) are the phases where AI isn't deployed. The studio has invested its AI capability in craft excellence. The next investment should be in business intelligence.
The 80/20 of Automation
Three automations would recover the most time: (1) Repurposing cascade — triggered automatically when a piece reaches "Published" status. (2) Distribution checklist — per content type, auto-generated, tracked in Dart. (3) Monthly performance report — AI agent collects metrics from LinkedIn + website + email, generates summary. These three alone would make Phases 5-7 operational instead of empty.