An executive sponsor announces AI priorities but repeatedly rewards teams only for the number of prototypes created. What behavior is likely to improve alignment?

1. Align incentives and reviews with measured business outcomes, responsible use, and production learning.
2. Increase the prototype quota while keeping outcomes unmeasured.
3. Reward only the largest model selected.
4. Stop all portfolio reviews after the announcement.