Stop Building AI Agents. Build AI Employees Instead (Live Demo) | Pedro Franceschi
Pedro Franceschi, CEO of Brex, joins Peter Yang to explain why founders should build AI employees with defined jobs rather than open-ended agents, and how to run them safely at scale.
The gist
-
Build agents with specific, limited jobs like a recruiter or analyst rather than general-purpose agents that can do anything.
-
Network-level security, not just prompt instructions, is needed to stop agents from crossing boundaries in production.
-
Track token spend by use case, model, and customer so you can find and fix cost problems precisely.
-
When hiring engineers today, test how well they build with AI tools, since proficiency can make them roughly 10x more productive.
"This is Jim. Jim is our OpenClaw recruiter. It runs on effectively like a three-stage pipeline. We do use OpenClaw for this but the harness doesn't really matter as much as just the skills. I have this system that I call autopilot and basically what it does is it's an OpenClaw..."
▶ Play at 12:3830 creators recommend it, including Claire Vo, This Week in Startups, and Garry Tan.
See all on the OpenClaw pageWhat they recommend · 1 pick
"This is Jim. Jim is our OpenClaw recruiter. It runs on effectively like a three-stage pipeline. We do use OpenClaw for this but the harness doesn't really matter as much as just..."
Peter, Riley, Claire and Liam agree on 39 things
See what they all use, in their own words. Paid ads are left out.
See what they all swear by Or pick your own peopleAlso mentioned
Ads in this episode
Riverside. They're paid, so we keep them out of the picks above.
See what Peter recommends without being paid →Related episodes
Episodes covering similar topics
More from Peter Yang