AI Strategy and Governance
AI Strategy and Governance
This domain tests whether you can connect AI ambition to enterprise strategy and install governance that enables rather than blocks delivery. Expect questions on portfolio prioritization, decision rights, policy hierarchy, and risk appetite.
Core rules and concepts
- Strategy cascade: start from enterprise objectives, then define AI value themes, then initiatives; avoid starting from vendor tools.
- Governance design: an AI council or board sets policy and risk appetite; delivery teams own outcomes; a central team owns standards, reuse, and assurance. Document decision rights with a RACI or similar.
- Policy hierarchy: principles to standards to procedures to controls. Policies should cover data use, model risk, human oversight, procurement, and incident escalation.
- Portfolio scoring: rank use cases on value, feasibility, data readiness, and risk; fund a balanced portfolio, not only quick wins.
- Governance gates: intake, design review, pre-deployment, and post-deployment review, proportionate to risk.
Worked example
A manufacturer has 40 AI ideas. It maps each to one of three strategic themes: uptime, yield, and safety. It scores value (annual benefit), feasibility (data and skills), and risk (safety, regulatory). A safety-critical vision inspection scores high value but high risk; it gets a staged gate with independent validation. A low-risk procurement chatbot scores lower value but quick payback; it is funded as a quick win. Reasoning: governance allocates scarce attention by risk and strategic fit, not by enthusiasm.
Traps
Treating governance as a sign-off queue; conflating strategy with tool selection; no executive owner; funding pilots without a route to scale; and writing principles with no measurable controls.