Context
1. Frame the operational problem
Agree the decision, delay, risk, or repeated work in scope. State what is excluded and who owns the outcome.
Context
2. Observe and baseline
Map normal work, workarounds, systems, data, approvals, exceptions, and representative performance. Record definition and confidence beside each baseline.
Context
3. Prioritize and design
Simplify before automating. Compare options against value, effort, risk, data readiness, adoption, continuity, and operating cost.
Action
4. Run a bounded pilot
Use representative data and users. Test failure, security, fallback, reconciliation, and support—not only the happy path.
Action
5. Roll out deliberately
Release by workflow, site, product, or user group where practical. Keep acceptance gates, communications, training, cutover, and rollback explicit.
Action
6. Hand over and improve
Transfer ownership, runbooks, access, monitoring, vendor dependencies, change process, and measurement reviews. Treat new evidence as an input to the roadmap.
