Overmind, a startup building a supervision layer for AI agents, has raised £2 million in a seed round led by Osney Capital, with participation from 14Peaks, Portfolio Ventures, Antler and Endurance Ventures. It develops technology designed to observe, secure and improve the performance of agentic AI systems. The funding will be used to grow technical teams, accelerate product development and scale go-to-market operations within industries including legal, healthcare and fintech.
As advances in agentic AI accelerate, the threat landscape is undergoing a shift that existing security tools are ill-equipped to handle. According to Gartner, 40% of agentic AI projects will be cancelled by 2027, driven in large part by inadequate risk controls. These models are vulnerable to adversarial inputs and data corruption, while risks also arise from how agents behave when executing tasks in live environments.
Overmind’s platform provides visibility into agent behaviour, detecting and preventing deviations in real time before humans could react. Through reinforcement learning, it improves agent performance and accuracy over time.
The founding team combines experience from the UK intelligence community and high-growth technology companies. Tyler Edwards previously spent eight years building AI systems for British intelligence agencies. Akhat Rakishev previously led machine learning infrastructure at Monzo and Lyst.
