---
title: "Interview agenda D — Hiring, governance and AI"
slug: "expert-q-a/interview-agenda-d-hiring-governance-and-ai"
section: "Expert Q&A"
description: "Potential experts: an experienced startup recruiting/people operator; a VC board professional; an AI risk practitioner who can interpret NIST's published v"
internal_only: false
---

# Interview agenda D — Hiring, governance and AI

**Potential experts:** an experienced startup recruiting/people operator; a VC board professional; an AI-risk practitioner who can interpret NIST's published voluntary profile. [Reboot podcast: How To Be a Stable Board Member](https://www.reboot.io/episode/164-how-to-be-a-stable-board-member-with-aileen-lee-brad-feld/) [NIST Generative AI Profile, AI RMF](https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-generative-artificial-intelligence)

### Q10. What proves a startup needs a full-time VP Sales rather than a better sales process?

**Why buyers care:** The wage and onboarding commitment may be hard to reverse. **Good answer covers:** founder-led motion, pipeline quality, stage conversion, repeatability, current managerial bottleneck, alternatives and milestone-based probation. **Expert quote:** *[Pending interview]*.

### Q11. How can a board challenge the founder without letting investor portfolio incentives become the answer?

**Why buyers care:** Directors can possess insight and conflicts simultaneously. **Good answer covers:** good-faith governance, decision authority, recusal where applicable, documented dissent and appropriate independent advice. Reboot's board episode discusses stability and blind spots, without proving any particular legal position. [Reboot podcast: How To Be a Stable Board Member](https://www.reboot.io/episode/164-how-to-be-a-stable-board-member-with-aileen-lee-brad-feld/) **Expert quote:** *[Pending interview]*.

### Q12. What evidence would convince you that an AI-generated pre-mortem is trustworthy enough to use?

**Why buyers care:** Polished text is cheap, verified judgement is not. **Good answer covers:** provenance, test cases, known failure modes, omitted options, versioning, data retention, human sign-off, confidential input and materiality thresholds. [NIST Generative AI Profile, AI RMF](https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-generative-artificial-intelligence) **Expert quote:** *[Pending interview]*.