The Pulse
Fleuret Raises €4 Million for AI-Powered Penetration Testing
The Paris-based startup says it will use its €4 million pre-seed round to develop AI agents that test systems for vulnerabilities and verify fixes. Fleuret describes a move from occasional penetration tests to ongoing testing as customer sy
AI.info Team · Written with AI ·
Fleuret AI has raised €4 million in pre-seed funding to build a platform that uses AI agents to test company systems for vulnerabilities and check whether fixes work. The Paris-based startup announced the round on October 5, 2026, with RAISE Ventures leading and Auriga Cyber Ventures, Wind Capital, Better Angle and cybersecurity-sector business angels also participating.
From periodic audits to repeat tests
Fleuret is pitching a shift from penetration tests conducted as occasional audits to security testing that can repeat as a company’s systems change. The company argues that a test can become outdated soon after it is completed: a new deployment may alter the systems exposed to attack. It also points to NIS2, the European cybersecurity directive, as a source of growing pressure on organisations to demonstrate their security practices.
The platform is organised around five tasks: mapping exposed systems, finding vulnerabilities, checking whether they can be exploited, helping teams fix them and retesting to verify the fix. Fleuret says its two agents, Emile and Champollion, inspect applications, APIs and infrastructure, then attempt to exploit vulnerabilities they identify. The company says each finding includes proof of compromise rather than a report based only on a possible weakness.
What Fleuret says its agents can do
After testing, the platform is intended to keep monitoring a company’s attack surface and trigger further tests when systems change. Fleuret is also developing prioritisation based on exploitability, links to tools used by engineering teams, automatic retesting and reports confirming whether a vulnerability has been fixed. The announcement describes the planned workflow; it does not provide independent performance results or a measured comparison with human-led penetration tests.
Yanis Grigy, Fleuret’s CEO and co-founder, said:
“We do not simply want to automate penetration testing as it exists today. Our ambition is to build the offensive security layer that will keep companies secure all year round, whatever their size. This round gives us the means to accelerate that vision and to build a platform designed for enterprise environments from day one.”
The company says the funding will support hiring in AI, software engineering and offensive security, as well as faster platform development. Fleuret says it aims to make offensive security accessible to every company, including those that cannot afford regular penetration tests.
Around ten people and customers including Brevo
Fleuret says it has around ten people and several customers, including Brevo, Stoïk and Yogosha. Its founders, Grigy and CTO Augustin Ponsin, previously started a company focused on penetration testing. The announcement says they have known each other since childhood, graduated from Télécom Paris and EPITA, and joined the X-HEC Entrepreneurs programme. Fleuret says the service operates from Europe on European infrastructure, a positioning aimed at organisations with data-sovereignty requirements.
The funding announcement does not disclose a valuation, the size of individual investments or a timetable for the planned hires. The practical test for Fleuret will be whether its agents can repeatedly find exploitable weaknesses, provide evidence teams can act on and confirm that repairs hold as customer systems continue to change.