Legal
AI use policy.
Draft. Pending counsel review.
- AI handles the repeatable work: readiness paths, application pre-fill, investor and mentor shortlists, first drafts of documents and decks, status updates.
- People make the decisions that carry risk: acceptance, pricing, equity, term sheets, and what gets launched or raised.
- Founder data is not used to train shared AI models.
Validate
Build
Launch
Scale
Capital
AI drafts
Draft
Draft
Draft
Draft
Draft
People decide
Decide
Decide
Decide
Decide
Decide
You approve
Approve
Approve
Approve
Approve
Approve
| Stage | What AI does | What people do | What you control |
|---|---|---|---|
| Validate | Builds your readiness path. Drafts interview guides and summaries. Prepares first-draft research. | Review interview findings with you. Challenge the business model. Make the go/no-go recommendation. | Edit, accept, or reject every output. |
| Build | Drafts requirements and user stories. Generates first-pass prototype screens. Prepares documentation drafts. | Approve scope and architecture. Review code and security. Decide what ships. | Edit, accept, or reject every output. |
| Launch | Drafts messaging options. Builds target account shortlists. Prepares outreach first drafts. | Set pricing with you. Approve launch messaging. Open customer conversations. | Edit, accept, or reject every output. |
| Scale | Summarizes pipeline and operating data. Drafts board update outlines. Prepares job description drafts. | Coach your sales and leadership team. Review hiring decisions. Advise on board matters. | Edit, accept, or reject every output. |
| Capital | Drafts the deck and data room index. Builds investor shortlists by thesis. Checks model formulas for errors. | Pressure-test the story and numbers. Review term sheets with you. Make introductions you approve. | Edit, accept, or reject every output. |
Accuracy rules
- Numbers come from calculations, not from AI text.
- Every AI output shows its sources.
- Human approvals are recorded.
- Low-confidence outputs go to a person.