The model
Two axes. One AI lens.
The practice is a matrix, not a list. Method is the journey an organisation moves through. Domains are the pillars you look at. Read any cell as the intersection — "Business Architecture × Roadmap" is sequencing AI into your operating model without breaking accountability.
Axis 1 — Method
The journey: six stages.
Baseline
Establish the current-state AI estate — what is running, shadow AI, exposure.
Ambition
Define target state and AI posture: cautious adopter through to AI-native.
Strategy
Choose where AI creates value — and where it must not go.
Roadmap
Sequence the moves without breaking accountability or budget.
Govern
Risk classification, review boards, guardrails, assurance — the spine.
Integrate & Operate
Embed into existing people, process and tools; run it.
Axis 2 — Domains
The pillars: five lenses and a spine.
Business Capability
Which capabilities become AI-augmented vs AI-native; opportunity-vs-maturity heatmap.
Business Architecture
AI in value streams; human/AI hand-offs; target operating model; decision rights; accountable officials.
Application Architecture
The AI application portfolio; assistive vs agentic; embed/build/buy; copilot / RAG / MCP / agent patterns.
Data & AI
The data foundation, model layer, vector/feature stores, LLMOps, the reference architecture itself.
Technology
Platform & compute; sovereign vs hyperscaler AI; AI-native security.
Governance
Cross-cuts every domain above; also scored as its own axis on the maturity radar.
Axis 1 × Axis 2
The artefact matrix.
Each cell is a candidate artefact. ◆ marks a launch anchor — done to authority standard first. The rest is a deliberate backlog: we do not build all thirty up front.
| Method ↓ / Domain → | Business Capability | Business Architecture | Application Architecture | Data & AI | Technology |
|---|---|---|---|---|---|
| 1. Baseline | AI capability heatmap | Shadow-AI discovery | AI app portfolio inventory | Data readiness audit | Compute / platform stocktake |
| 2. Ambition | AI ambition canvas | Target operating model options | Assistive → agentic posture | Data product vision | Sovereign-AI stance |
| 3. Strategy | Value vs risk prioritisation | Decision-rights model | Build / buy / embed policy | Data-sharing & provenance strategy | Reference-architecture choice |
| 4. Roadmap | Capability sequencing | Operating-model transition | Portfolio roadmap | Data-foundation roadmap | Platform roadmap |
| 5. Govern | Capability-level risk owners | Accountable-officials map | App-tier review gates | Provenance evidence for AIA | AI-native security controls |
| 6. Integrate & Operate | Workforce / skills uplift | Process redesign with AI | LLMOps & lifecycle | Monitoring, drift, lineage | Run-cost & FinOps for AI |
Governance cross-cuts every cell above and is also scored as its own axis on the maturity radar.