Most teams do not need another broad conversation about whether AI is good or bad. They need to decide where it helps, where a person must stay accountable, and what to try next without turning every workflow into an experiment.
These six AI at work templates are built for that moment. Each one starts with a real team decision and ends with an owner, a small next step, and something you can review later.
Choose the template that matches your decision
1. Review a real workflow before automating it
Use the AI Workflow Retrospective when a process already includes AI—or when people are quietly using it in different ways. Map what became faster, where quality slipped, what new review work appeared, and which guardrail is missing. It is useful for product, engineering, support, operations, and any team tired of comparing anecdotes.
2. Decide what AI can do and what people must own
The Human–AI Delegation Map makes responsibility explicit. Sort work into AI can draft, AI can recommend, a human must decide, and a human must do. Then name the person accountable for the result. Use it before adopting an agent, automating a handoff, or changing a policy.
3. Test product discovery evidence, not AI enthusiasm
Use the AI Product Discovery Review to examine a customer problem, the evidence behind it, AI's possible contribution, and the riskiest assumption. Product managers can use it to stop “add AI” from becoming the strategy and find the smallest honest test instead.
4. Make an AI feature decision people can revisit
The AI Feature Decision brings customer value, model limits, data needs, failure modes, and operating cost into one decision. The output is not a forever promise. It is a clear go, test, narrow, or stop decision with reasons the next team can understand.
5. Map how roles and skills are changing
Use the AI Role Impact & Skills Map with HR, people leaders, and team leads. Identify tasks likely to shrink, grow, or change; the judgment that still matters; and the skill people need next. This turns workforce planning into a conversation with the team, not a spreadsheet about them.
6. Surface trust and safety concerns before they become silence
The AI Trust & Safety Check gives people a structured way to raise concerns about privacy, bias, explainability, reliability, and psychological safety. Use it before a rollout and again after real use. A concern does not automatically block progress; it tells you which condition must be true for progress to be responsible.
What each role gets from the templates
- Product managers: stronger discovery evidence, explicit trade-offs, and smaller AI bets.
- HR and people teams: a grounded view of role change, skills, trust, and employee voice.
- Team leads: clearer ownership, safer delegation, and a practical experiment for the next sprint.
- Teams: permission to name hidden work and uncertainty without being labelled resistant to change.
Do not stop at the workshop
A good board creates clarity. Behavior change needs a return loop. Once the team chooses an experiment, put it into Change Loop, collect lightweight signals during the work, and review whether to keep, adapt, or stop it at the next retro.
Start with the decision that is already causing friction. Pick one template, invite the people closest to the work, and leave with one small move you can actually learn from.