Before launch, walk through a realistic bad day with the people who'd be involved. It takes an hour and it surfaces gaps that no amount of design review will.

Prompt: Running an Incident Rehearsal

Help me rehearse an incident for an agent we're about to launch, so we find the gaps before they find us.

THE AGENT: [what it does]
ITS TOOLS: [list, read/write marked]
APPROVALS: [what requires a human yes]
VOLUME: [tasks per day]
WHO WOULD BE ON CALL: [roles]

Please:
1. Write three realistic incident scenarios, in increasing severity. Make them specific and plausible given my tools, not generic AI-gone-wrong stories.
2. For the worst one, walk through the timeline hour by hour: when it starts, when someone notices, what they see first, what they do.
3. At each point in that timeline, name what we'd need in place to detect or stop it, and ask me whether we have it.
4. What's the single cheapest change that would most reduce the damage in this scenario?
5. Write the one-page runbook: how to stop the agent, who can do it, how to work out what it did, and how to reverse what's reversible. Assume the reader is stressed and it's late.

Agentic AI and MCP for Product Managers

Module 1What 'Agentic' Actually Means

Module 2How an Agent Actually Works

Module 3MCP: What It Is and Why You Should Care

Module 4Designing the Tool Layer

Module 5Permissions and the Human in the Loop

Module 6Evaluating Agents

Module 7Cost, Latency and Unpredictability

Module 8Designing What the User Sees

Module 9Risk and Blast Radius

Module 10When an Agent Is the Wrong Answer

Module 11Final Project: Design One Real Agent End to End

Module 12Appendix: The Prompt Library

Rehearse the Incident