Marketplace Tech · Thursday, September 17, 2026
Inez Boumajaja's startup, Clemy, is developing a method to quantify AI risk by testing AI agents for potential errors before they can cause claims. By simulating errors like leaking customer data, Clemy aims to provide a 'score' that correlates with lower insurance premiums. This approach is likened to ensuring safety for new infrastructure like skyscrapers.
“How do you know if a brand new agentic AI system your company uses to process invoices is likely to make a costly mistake? That's what Inez Boumajaja is trying to quantify with her AI insurance startup, Clemy.”
“When we onboard a customer, what we actually do is we test their AI agents. We try and actually trigger the potential errors that would lead to a claim. Like leaking sensitive customer data.”
“We had the agent in under like, 10 minutes, give you information about the previous customer they talked to, their age, your, you know, like personal information.”
“The better the score, the lower the insurance premium. We pride ourselves ourselves of being, you know, the enabler of technology adoption. It's like there wouldn't be any skyscrapers in New York if you didn't have like the proper insurance for it.”