Guide
How to use founder playbooks with AI agents
A practical guide for turning founder playbooks into source-grounded prompts, checklists, decision memos, and AI-agent workflows.
Last updated
Published by ReScience Lab Inc., maker of NoIdea.
Start with the decision
Name the decision before prompting the agent: fundraising prep, GTM channel selection, hiring process design, customer discovery, or product prioritization.
Extract context before asking for output
Pull out constraints, assumptions, examples, failure modes, definitions, and seller caveats from the playbook. This reduces generic answers and makes the agent easier to audit.
Ask for inspectable artifacts
Prefer checklists, comparison tables, risk registers, interview questions, experiment plans, and decision memos over vague final recommendations.
Agent prompt patterns
| Goal | Ask the agent for | Avoid |
|---|---|---|
| Prepare | Questions and missing context | A guaranteed answer |
| Compare | Criteria table and tradeoffs | One-size-fits-all advice |
| Decide | Decision memo with caveats | Legal/financial advice framing |
FAQ
Should an AI agent make the final startup decision?
No. The agent can organize context and questions, but founders should make decisions with judgment and appropriate professional advice where needed.
What output is safest to request?
Ask for questions, checklists, risks, comparison tables, or decision memos that a human can inspect.
Related NoIdea pages
The core product page for NoIdea marketplace assets.
Context pack checkerCheck whether an asset is ready for agent workflows.
Knowledge asset glossaryA definition for packaged founder/operator expertise.
Setup guideMachine-readable setup notes for CLI and agent workflows.