​Vercel trained an AI agent on its best salesperson—then used it to shrink a 10-person inbound team down to one human in the loop.

​Drew Bredvick is the Director of GTM Engineering at Vercel and the person who helped design, ship, and operationalize that system.

​In this session, Drew walks us through the actual mechanics behind Vercel’s agents: how they’re trained on real workflows, how tasks are broken down and routed, where humans stay in the loop, and how quality and trust are enforced at scale.

​We’ll get specific. You’ll see how inbound messages are classified, how agents qualify leads, draft responses, hand off edge cases, and improve over time—plus how Vercel decides which work is safe to automate and which work stays human.

​You’ll leave with a clear mental model—and a repeatable pattern—for building AI agents that take on high-volume, entry-level work across sales and ops without slowing teams down.

​(Engineer-friendly.)