Organization, Talent, and Change Management
Organization, Talent, and Change Management
This domain tests the operating model, roles, skills, and adoption practices that turn AI capability into routine work. Expect questions on centralization versus federation, role design, workforce transition, and adoption metrics.
Core rules and concepts
- Operating models: centralized AI center of excellence, federated embedded teams, or hub-and-spoke. Choose based on scale, risk, and business-unit maturity; hub-and-spoke often balances standards with domain ownership.
- Define roles clearly: AI product owner, ML engineer, data engineer, data steward, risk or compliance partner, and business process owner. Avoid orphaned accountability.
- Skills taxonomy: data literacy for managers, AI literacy for product owners, deep technical skills for builders, and risk skills for reviewers. Develop career paths, not one-off training.
- Change management: identify affected workflows, incentives, and fears. Use sponsors, champions, and feedback loops. Adoption is a behavior change, not a launch event.
- Measure adoption: active use, override rates, time saved, training completion, and support tickets. Pair adoption metrics with value metrics.
Worked example
A bank starts with a central AI team. Business units complain about slow delivery. The bank moves to hub-and-spoke: the hub owns platforms, standards, and high-risk assurance; spokes own domain use cases and data products. It adds AI product owners in each spoke and a shared skills curriculum. Reasoning: the model change follows a delivery bottleneck and risk need, not a fashion; accountability stays with the business.
Traps
Leading with tools instead of workflow change; no incentives for adoption; underestimating data literacy; and centralizing all decisions so business owners disengage.