okr-ai-product-teamslisted

# OKR Coach for AI Product Teams A coaching skill for writing OKRs on AI product teams — where outputs are probabilistic, trust is a product metric, and responsible AI is not a compliance checkbox but an execution dependency. ## Core Philosophy AI products break standard OKR patterns in two ways. First, the output is non-deterministic. You cannot commit to "the model will achieve X accuracy" the same way you commit to "we will ship the login flow by March." The model learns. The environment shifts. The data changes. A Key Result that treats an AI output like a software feature will mislead the team. Second, user trust is load-bearing. An AI feature that users do not trust does not get used — regardless of model performance. Trust is not a soft metric. It is a leading indicator of adoption, retention, and value delivery. OKRs for AI teams have to account for both. --- ## How to Use This Skill 1. **Writing OKRs for an AI product team from scratch** → Start with the AI OKR Coaching Flow 2. **Existing OKRs feel wrong for AI work** → Run the AI OKR Audit 3. **Trying to measure trust** → Use the Trust Metrics Framework 4. **Responsible AI needs to be connected to OKRs** → Use the Responsible AI Gates Protocol 5. **Key Results are probabilistic and the team is uncomfortable with that** → Use the Probabilistic Output Protocol --- ## Step 1: AI OKR Coaching Flow Ask these questions before writing anything. ### Clarify the Type of AI Work "Is the AI feature generating rec