# Team AI Onboarding
Practical playbook for team leads introducing AI to their team. Not enterprise governance. Hands-on team adoption.
## Step 1: Assess Team AI Literacy
Who's using what already? Most teams have [[Shadow AI]] happening whether you know it or not. Survey the team: What tools are people using? For what tasks? What data are they putting in? This is your baseline. See [[AI Literacy]].
## Step 2: Define Team AI Policy
What tools are approved. What data can be shared. What requires review before shipping. This doesn't need to be a 50-page document. One page covering approved tools, data boundaries, and review requirements. See [[AI Usage Policy]].
## Step 3: Create Shared Context
Team CLAUDE.md (or equivalent) with coding conventions, architectural decisions, processes, priorities. This is the single highest-leverage action. Without shared context, every team member reinvents the wheel. See [[Team Context Management (TCM)]] and [[Context File Hierarchy]].
## Step 4: Start With One Shared Skill
A team code review skill. A PR template skill. A deployment checklist skill. Pick one workflow the whole team does, codify it, and share it. This demonstrates the value of [[AI Skill Distribution]] in concrete terms.
## Step 5: Establish Review Practices
Never ship unreviewed AI output. Code review still applies. Content review still applies. AI is a draft generator, not a final answer. See [[Human-in-the-Loop]].
## Step 6: Measure and Iterate
What improved? What broke? What do people need? Run a retrospective after 2-4 weeks. Adjust the policy, update the context, refine the skills.
## Common Failure Modes
- **Forcing adoption.** Let early adopters lead. Mandating AI use creates resentment and low-quality adoption. The skeptics come around when they see results.
- **No shared context.** Everyone reinvents. Five people on the same team writing five different system prompts for the same codebase. Waste of time.
- **No quality gate.** AI output ships unchecked. This erodes trust fast. One bad AI-generated bug in production sets adoption back months.
- **Tool proliferation.** Standardize on 1-2 tools. Every new tool is a new context silo, a new security surface, a new thing to learn. See [[AI Agent Permissions]].
## References
## Related
- [[Team Context Management (TCM)]]
- [[AI Usage Policy]]
- [[Shadow AI]]
- [[AI Skill Distribution]]
- [[AI Agent Permissions]]
- [[Human-in-the-Loop]]
- [[Context File Hierarchy]]
- [[AI Literacy]]