02 - The Roadmap
Three tracks, three tempos - concurrent, not sequential
flowchart TB A["TRACK A - Culture & adoption (people)<br/>tempo: continuous, human-paced, never done<br/>= organization maturity"] B["TRACK B - Infrastructure & system (substrate)<br/>tempo: slow, phased, built in step with adoption<br/>= data maturity + knowledge backbone"] C["TRACK C - Product & workflow (applications)<br/>tempo: fast, ship-test-iterate<br/>= the AI surfaces, the compounding engine"] A -->|"paces how fast workflows hand off"| C B -->|"gates what each workflow can do"| C classDef c fill:#C6EFCE,stroke:#2E7D32,color:#111; class C c;
Tracks A and B produce the shared knowledge layer; Track C consumes it. All three move at once.
The sequencing rule
Track C never waits for Track B to be “finished.” Each workflow ships the moment the specific substrate rung it needs exists. Order Track C by least substrate required -> most, banking compounding value early while A and B grind underneath.
flowchart LR R1["Reporting Tier 1<br/>needs: ads MCP only"] --> ID["Ideation<br/>needs: vault + scraping"] --> CR["Creative production<br/>needs: angle library + render"] --> SO["Social mgmt<br/>needs: governance + Zernio"] --> DM["DM / CS bot<br/>needs: solid knowledge base"] --> R2["Reporting Tier 2<br/>needs: the warehouse"] classDef first fill:#FFF2CC,stroke:#B7950B,color:#111; class R1 first;
Two bars on every Track C item: shipped (it works, value proven) and handed off (a colleague runs it without you). Ship at product-speed; hand off at culture-speed. This is what stops every workflow becoming a bottleneck around one person.
Phased plan
gantt title EFFEN Applied AI - phased plan dateFormat YYYY-MM-DD axisFormat %b section Track A - Culture Bench + onboarding path :a1, 2026-07-01, 60d WhatsApp -> Notion/Workspace :a2, 2026-08-01, 120d Train + support operators :a3, 2026-09-01, 150d section Track B - Infra Knowledge backbone + tokens :b1, 2026-07-01, 45d Warehouse phase 1 (1 brand) :b2, 2026-08-15, 45d Warehouse fan-out + US :b3, 2026-11-01, 120d section Track C - Product Reporting Tier 1 :c1, 2026-07-08, 20d Ideation :c2, 2026-08-01, 40d Creative production :c3, 2026-09-01, 60d Social management :c4, 2026-10-15, 45d DM / CS pilot :c5, 2026-10-01, 60d Reporting Tier 2 (MER) :c6, 2027-01-01, 60d
Phases are waves, not gates. Success signals: someone pulls ad performance with zero manual export (P0); first agent-drafted brief cleared + one live dashboard (P1); operators run creatives in their own accounts + approved calendar publishing (P2); MER-based decisions replacing last-click reads (P3).
Dependency map - what gates what
flowchart TD COWORK["Cowork + ads MCP"] --> R1["Reporting Tier 1"] VAULT["Vault + scraping + Foreplay"] --> ID["Ideation"] ID --> ANG["Angle library + performance ledger"] ANG --> CR["Creative production"] RENDER["Canva/Creatomate render + AI image gen"] --> CR GOV["Vault governance + Zernio + approval"] --> SO["Social management"] KB["Solid knowledge base (OpenKnowledge)"] --> DM["DM / CS bot"] META["Meta app review + business accounts"] --> DM WH["The warehouse (fan-out)"] --> R2["Reporting Tier 2"] classDef first fill:#FFF2CC,stroke:#B7950B,color:#111; class R1 first;
| Product item | Requires | Ships |
|---|---|---|
| Reporting Tier 1 | Cowork + ads MCP only | Phase 0 (now) |
| Ideation | Vault + scraping + Foreplay + angle library | Phase 1 |
| Creative production | Raw materials + angle library + ledger + render | Phase 1-2 |
| Social management | Vault governance + Zernio + approval | Phase 2 |
| DM / CS bot | Knowledge base + Meta app review | Phase 2-3 |
| Reporting Tier 2 | The warehouse (fan-out) | Phase 3 |
| Video / marketplace / CRM | Own tooling + data | Phase 3+ |