No segment of cybersecurity is immune to the impact of AI.
With a €4 million pre-seed funding round, Fleuret AI aims to move pentesting, which is still often conducted as a point-in-time audit, toward a continuous security approach.
For the startup, a classic intrusion test only provides a snapshot of a system’s security at a given moment, while a new deployment can quickly alter a company’s attack surface.
To this end, it has developed two AI agents—Emile and Champollion—that map a company’s environment, explore its applications, APIs and infrastructure, and then attempt to exploit the vulnerabilities identified.
Each detected flaw is accompanied by a proof of compromise, intended to concretely demonstrate its exploitability. The platform then continuously monitors the attack surface and can relaunch tests as the system evolves.
Its platform rests on five pillars: knowing exposed systems, identifying vulnerabilities, demonstrating their exploitability, facilitating their remediation and verifying their effective resolution. It aims in particular to help organizations meet strengthened security requirements, including those associated with the European NIS2 directive.
Founded in 2026 in Paris by Yanis Grigy, CEO, and Augustin Ponsin, CTO, Fleuret AI now counts around ten employees and several clients, including Stoïk, Ouwba, Brevo and Yogosha.
Its funding round was led by RAISE Ventures, alongside Auriga Cyber Ventures, Wind, Better Angle, and cybersecurity industry angels.
“Our ambition is to build the offensive cybersecurity layer that will enable companies to stay secure year-round, regardless of their size,” says Yanis Grigy, CEO and cofounder of Fleuret AI.
