Fleuret AI raises €4 million for continuous penetration testing

Fleuret AI has raised €4 million to develop a platform that tests company systems for security weaknesses on an ongoing basis. The French startup says it will use the pre-seed funding to expand its platform and recruit specialists in AI, software engineering and offensive cybersecurity. The round was led by RAISE Ventures, with Auriga Cyber Ventures and other investors participating.

The funding concerns a commercial cybersecurity product, not a regulatory decision or a change in legal status. Fleuret AI says its platform is designed to make penetration testing continuous, while conventional tests provide a snapshot of systems at one point in time. A penetration test is a controlled attempt to find and demonstrate exploitable weaknesses in software or infrastructure.

The platform checks changing systems

Fleuret AI says its platform covers five stages: mapping exposed systems, finding vulnerabilities, testing whether they can be exploited, supporting fixes and checking that those fixes work. These steps are intended to connect discovery with remediation, the process of addressing identified security issues. The company describes the service as a continuing process rather than an isolated audit.

Its two AI agents, Emile and Champollion, map a company's environment and examine its applications, APIs and infrastructure. According to the source material, the agents then attempt to exploit vulnerabilities they identify. Each finding is accompanied by proof of compromise, which is intended to show whether the weakness can be exploited.

Founders set out the funding plan

Fleuret AI was founded in 2026 by Yanis Grigy, its CEO, and Augustin Ponsin, its CTO. The company's stated use for the funding is to accelerate platform development and grow its team in AI, software engineering and offensive cybersecurity. No hiring target or allocation of the €4 million among these activities was provided.

The company also describes tools under development to help technical teams address findings. These include prioritising vulnerabilities according to exploitability, connecting with teams' existing tools, automatically testing systems again and producing reports on whether vulnerabilities have been fixed. The source material does not specify when these tools will become available.

Fleuret AI says its agents operate from Europe on European infrastructure, a design it associates with sovereignty requirements in highly regulated sectors. The company is targeting the European market and says it wants its platform to serve organisations that may lack resources for regular penetration testing. No independent assessment of the platform's performance or security outcomes was included in the material.

The funding announcement names RAISE Ventures as lead investor and Auriga Cyber Ventures among participants. The source material also reports participation by Wind, Better Angle and cybersecurity business angels; a separate account specifies Wind Capital. It does not provide individual investment amounts, a company valuation or further terms of the pre-seed round.

The company frames continuous testing as a response to systems changing after an audit, which can alter the attack surface. Its platform is intended to monitor that surface and start further tests when systems change. The source material does not describe how often checks run, what access customers must provide or how customers approve attempted exploitation.

The available accounts set out the product's intended functions and the planned use of funding, but give no deployment figures or customer names. They also do not state a launch timetable for the platform's developing remediation features. Fleuret AI's next reported steps are team expansion and faster platform development, with no milestones specified. Further information would be needed to assess product availability, customer adoption and the results of continuous testing.