A transformation roadmap assumes all departments have the same AI maturity. In practice, some lack reliable data and others already operate models. What is the better approach?

1. Assess capabilities by context and tailor phased enablement while sharing useful practices.
2. Keep all experienced teams isolated so their learning cannot spread.
3. Force identical deadlines and controls without assessing any differences.
4. Treat model access alone as a complete maturity assessment.