Bright Signal - European Tech Founders & Expert Interviews podcast

Maurits Kaptein from Kyvvu: when should an AI agent ask permission?

0:00
1:11:22
Rewind 15 seconds
Fast Forward 15 seconds

In this episode, Bart & Raphaël sat down with Maurits Kaptein, co-founder and CTO of Kyvvu and full professor at TU Eindhoven, to unpack how you stop an AI agent from doing something it shouldn’t. Maurits explained why checking single actions is not enough, and how Kyvvu’s agent security kernel sits inside the agent’s harness and blocks risky sequences of actions before they happen.

We got concrete with a Dutch health insurer whose customer service agent must never combine someone’s claims with a sales pitch, talked partners, pricing per active agent and why the market is growing faster than any startup can keep up with, and dug into the €1M pre-seed round backed by Volta Ventures and BOM. We closed with Maurits’ advice for anyone letting an agent loose on their own computer.

🙋‍♂️ Guest: Maurits Kaptein, co-founder and CTO of Kyvvu  
📩 Connect on LinkedIn: https://www.linkedin.com/in/mauritskaptein/
🌐 Kyvvu: https://kyvvu.com

More episodes from "Bright Signal - European Tech Founders & Expert Interviews"