AI Marketing Department · Knowledge & governance
Machine for memory, human for meaning.
15 governed knowledge stores that every agent reads from and writes back to, and 20 points where a named person decides.
Shared knowledge
Who writes, who reads.
Written by · 0
Maintained by people: set up during onboarding (/sns-setup) and changed only through approved updates.
Knowledge store · IDs PRF-…
Proof Vault
Every record carries an ID, source, owner, status, confidence and expiry.
Used by 29 of 71 agents.
Read by · 29
- Battlecard Builder
- Brand Strategy Guardian
- Channel Activation Planner
- Competitive Response Agent
- Content Editor
- Content Gap Analyst
- Content Production Agent
- Content Strategy Agent
- Conversion Copy Agent
- Differentiation Analyst
- Email Copywriter
- Enablement Context Architect
- Knowledge Systems Manager
- Landing Page Strategist
- Launch Experiment Designer
- Launch Strategy Architect
- Lead Magnet Architect
- Longform Writer
- Marketing QA Agent
- Message Kernel Architect
- Nurture Sequence Architect
- Objection Handling Agent
- Offer Strategist
- On-Page SEO Optimizer
- Positioning Engine
- Sales Talk Track Agent
- Social Post Writer
- Thought Leadership Agent
- Value Proposition Architect
Reference
The fifteen stores
| Store | Record IDs | Written by | Read by |
|---|---|---|---|
| Company Context | CMP-… | People | 8 agents |
| Product Context | PRD-… | People | 9 agents |
| ICP Library | ICP-… | People | 18 agents |
| Buyer Personas | PER-… | People | 21 agents |
| Market Signals | SIG-… | 2 | 17 agents |
| Competitive Intelligence | CI-… | 1 | 13 agents |
| Positioning CanonCanon | POS-… | 1 | 25 agents |
| Brand Voice MemoryCanon | VOI-… | People | 19 agents |
| Message MemoryCanon | MSG-… | 1 | 31 agents |
| Proof Vault | PRF-… | People | 29 agents |
| Enablement Source of Truth (ESOT)Canon | ESOT-… | 2 | 10 agents |
| Campaign Experiments | EXP-… | 9 | 17 agents |
| Launch Learnings | LRN-… | 3 | 5 agents |
| Performance Observations | OBS-… | 8 | 24 agents |
| Decision History | DEC-… | 6 | 8 agents |
Human governance
Where people stay accountable.
- Human approval gates
- Canon changes, spend, sends, publishing and executive decisions stop for a named human approver, with written criteria, before the workflow can continue.
- Proof or it doesn't ship
- Decision-stage claims must cite an approved, unexpired ProofID from the Proof Vault. The check runs on every step.
- Every claim says what it is
- Outputs label each statement as hypothesis, AI recommendation, verified evidence, human decision, executed action or measured result.
- Governed memory
- A guard hook blocks knowledge edits that would put unapproved or unproven claims into canon. Every store records source, owner, status, confidence and expiry.
- Truth Pass and Taste Gate
- A truth pass and a human taste gate check tone, ethics and accuracy before anything customer-facing is approved.
- Insufficient data is an answer
- When evidence is thin, agents publish with an ‘insufficient data’ banner rather than padding.
- Outcome validation
- Narrative Fit, Engagement Quality, Sales Momentum, Launch Learning Rate and Enablement Latency are computed from logged observations, not asserted.
All 20 approval gates
- 1. Market PulseApprove implications and assign ownersHead of Product Marketing (or delegated PMM)
- 2. Positioning EngineApprove into positioning canonHead of Product Marketing (with CMO sign-off for category changes)
- 3. Messaging EngineApprove into message memoryHead of Product Marketing (brand lead for voice exceptions)
- 4. Learning LaunchBuild the Day −30 → +60 launch planLaunch owner (PMM lead) with budget holder for any spend
- 4. Learning LaunchDay 60 extend or evolve and write learning memoryHead of Product Marketing
- 5. Sales Enablement EngineOwner confirmation for pricing and roadmapPricing owner (pricing) or Product lead (roadmap); PMM for other topics
- 5. Sales Enablement EngineUpdate ESOT guidanceSales Enablement lead with PMM owner
- 6. Competitive ResponseBuild the 72-hour Counter KitHead of Product Marketing (legal review for comparative claims)
- 6. Competitive ResponseUpdate the battlecardSales Enablement lead
- 7. SEO Growth EngineOn-page optimizationContent lead (web owner publishes)
- 7. SEO Growth EnginePrioritize refresh and next clustersHead of Content / SEO lead
- 8. Demand Generation EngineChannel and distribution planDemand generation lead with budget holder
- 8. Demand Generation EngineDraft campaign contentContent lead (channel owners publish)
- 9. Lifecycle EngineStrategy and send approvalLifecycle/Email owner (sends from the ESP)
- 10. Conversion EnginePre-register the experimentGrowth/CRO lead (web owner launches the test)
- 10. Conversion EngineRecord validated learningGrowth/CRO lead
- 11. GTM Performance EngineApprove Stop / Start / ScaleCMO or VP Marketing (budget holder for Scale moves)
- 12. Executive Marketing BriefExecutive approval and decision logCMO (or executive sponsor)
- 13. Social & Community EnginePre-publish QA and human approvalSocial lead plus each named publisher (executives approve their own posts)
- 14. Editorial EngineHuman approval before publishingContent lead (offer owner for any commercial term)
The human team the book recommends
- System PMM
- Loop architect. Owns the Stack Canvas, data contracts and cadence.
- Story Director
- Taste and proof. Owns the Voice Card, the Canon and decision-stage proof.
- Field Orchestrator
- Moment of truth. Owns ESOT, copilot quality and the Objection Lab.
- CI Strategist
- Trajectory detection. Owns copy and pricing drift, counter kits and reframes.
- Analyst-in-residence
- Causal gym. Owns cohorts, counterfactuals and Command Board fidelity.
Team design from Systems, Not Slides, Chapter 10. Gate approvers are as named in the workflow specs.
AI Marketing Chief of Staff enforces every gate.