Bio
Three-time Chief Product Officer (LaunchDarkly, Optimizely, Color Genomics). Founder of ChatPRD, a tool for product managers to draft PRDs with an AI co-author. Hosts "How I AI", a podcast on operator-grade AI use. Vocal early skeptic of AI agents who reversed her position after running OpenClaw — a personal multi-agent setup with nine specialized agents on three Mac Minis covering work, sales, household, and family operations.
Operating themes
- Agents as a team, not a tool. Each role gets its own agent with its own context window and tool scope. Manage them like employees.
- Manager skill is the unlock. The bottleneck for productive AI agent use is operating discipline, not engineering depth.
- Progressive trust ramps. Stage tool access in named tiers, the way you onboard an executive assistant.
- Plain-language onboarding > structured forms. Voice notes and conversational ramps are higher-bandwidth than fill-in-the-blank UIs.
Cards
- Agents work when treated as a team, not a single super-tool — Don't throw every task at one agent; build a team [Tier A]
- The unlock for AI agent productivity is management skill, not technical skill — Operating discipline is the bottleneck, not engineering [Tier A]
- Onboard agents the way you onboard an EA: progressive trust, named tiers — Stage tool access the way you onboard an EA [Tier A]
Sources captured
- 2026-04-28 — Lenny's Podcast, "OpenClaw, agents as a team, manager skill as the unlock" (
raw/podcasts/claire-vo--openclaw-agents-as-team--2026-04-28.md)