Former MI5 officer lands €2.3 million for AI agent security startup Overmind

London-based Overmind, a start-up developing the guidance layer for AI representatives, today introduced a EUR2.3 million (₤ 2 million) Seed round to expand their technological groups, speed up item advancement and range go-to-market procedures within lawful, medical care and FinTech– where agentic AI holds prospective however needs strenuous governing conformity and information personal privacy.

The financial investment was led by expert cybersecurity financiers Osney Funding, with involvement from 14Peaks, Profile Ventures, Horn and Stamina Ventures.

Tyler Edwards, founder and chief executive officer of Overmind states: “ The AI safety and security sector is attempting to safeguard the incorrect point. Versions will certainly constantly be prone to adversarial inputs– that’s a basic residential or commercial property of exactly how they function. Yet what occurs when a representative is stay in manufacturing, engaging with genuine systems, and its behavior begins to wander? Today, many groups have no concept. Overmind gives the deployment-layer facilities required to check representative communications and step in prior to damages takes place.“

In the 2025– 2026 European financing landscape for agentic AI and AI security/governance innovations, a variety of start-ups have actually elevated funding that assists contextualise Overmind’s Seed round.

Archestra, a London-based business concentrated on securely attaching AI representatives to interior information, protected EUR2.8 million in pre-Seed financing to create security guardrails for self-governing systems. In Italy, Equixly elevated EUR10 million to scale its AI-driven API safety and security screening system. French start-up Qevlar AI elevated EUR9.1 million to construct agentic AI for safety and security procedures centres.

Various other nearby firms– such as Ranketta, with EUR1 million for AI analytics on brand name exposure in LLM outcomes, and Omnia, with EUR3.5 million for agent-driven advertising and marketing systems– additional show the circulation of funding right into innovations sustaining AI fostering a lot more extensively.

Taken with each other, these rounds stand for over EUR26 million in financing for start-ups dealing with either self-governing AI representatives or the safety and security and administration difficulties they bring.

Adam Cragg, Companion at Osney Funding: “ In the brand-new frontier of self-governing AI, representative safety and security, efficiency, and implementation are the utmost affordable benefits. Overmind gives organizations with really set apart modern technology that keeps track of and protects agentic AI while iteratively boosting version efficiency, allowing groups to range with self-confidence. We’re thrilled to back such a solid starting group dealing with an important market“

Established In 2025, Overmind is a devtool that allows AI to pick up from manufacturing information. Making use of pattern of life evaluation, it transforms real-world representative behavior right into constant renovation, supposedly enabling groups to deliver safe and very specialized representatives.

Overmind’s starting group incorporates knowledge from the UK knowledge neighborhood and high-growth modern technology firms.

chief executive officer Tyler Edwards invested 8 years developing AI systems for British knowledge companies consisting of MI5, MI6 and GCHQ. Akhat Rakishev, CTO, formerly led artificial intelligence facilities at Monzo and Lyst, while CRO Sam Force has actually scaled go-to-market at 3 unicorns: Financing Circle, Pipeline and Vertice.

Adam French, Companion at Horn, shares: “ Overmind is dealing with among one of the most vital traffic jams in the improvement of AI: the safety and security and guidance of self-governing representatives. The starting group is distinctively placed to fix this and provide the ‘intelligence-grade’ safety and security AI devices require. We are happy to back a group that isn’t simply developing a device, however is specifying the safety and security requirement for exactly how superintelligence will certainly be securely released in manufacturing“

As advancements in agentic AI speed up, the danger landscape is going through a basic change that existing safety and security devices are unfit to deal with.

The business discusses that these designs are naturally prone to adversarial inputs and information corruption, yet the genuine threats hinge on exactly how these representatives act when carrying out jobs in real-time atmospheres.

This safety and security space is ending up being an obstacle to development; Gartner approximates that 40% of agentic AI jobs will certainly be terminated by 2027, driven in big component by insufficient danger controls. Without unique modern technology, risky implementation will certainly paralyse the fostering and development of agentic AI.

Overmind is assisting organizations satisfy the assurance of agentic AI, with intelligence-grade modern technology that observes, protects and boosts representative efficiency. The system gives full exposure right into representative behavior, spotting and avoiding discrepancies in real-time prior to human beings might respond.

Moreover, via support understanding, Overmind supposedly boosts representative efficiency and precision with time, so they do not simply remain secure, they improve.

The message Former MI5 officer lands €2.3 million for AI agent security startup Overmind showed up initially on EU-Startups.