What RevOps should build vs. buy for the company brain (Ex-HubSpot CPO at Vector)

Get agents that know your customers better than you do: https://day.ai/ RevOps teams are building now. One person with Claude Code has the bandwidth of six engineers from 2017. So what should you build, and what should you still buy? At Vector 2026, Christopher O'Donnell (co-founder of Day AI, 10 years leading product at HubSpot) walks through what a company brain actually is, why traditional CRM is never in the top five data sources RevOps teams rely on, and how to decide part by part: meeting transcripts, data sources, the agent control plane, harnesses, work surfaces, knowledge management. His three questions to take back to your team: 1. What number are we moving with our AI strategy? 2. How do we get from baseline to 2x? 3. Who has control of what, and who can see what? Get agents that know your customers better than you do: https://day.ai/?utm_source=youtube&ut... 00:00 From HubSpot product to running RevOps 03:15 You are a product engineering team now 04:49 What a company brain actually is 05:53 Start with a folder 08:02 Brains scattered everywhere 11:42 What number are you moving? 15:31 What stays human, what goes to AI 19:12 Before: humans bot-sitting 21:18 After: proactive agents and self-improving loops 24:59 You + Claude Code = six engineers 25:31 What to build and what to buy 29:52 Three questions to take back to your team