In enterprise boardrooms today, technology leaders face a critical decision: how do you unlock AI agents' transformative potential while maintaining security and governance? From Silicon Valley startups to Fortune 500 companies, organizations are wrestling with autonomous AI adoption. A recent EY survey of 500 U.S. business leaders shows 9 in 10 recognizing agentic AI’s potential, yet struggling to overcome significant adoption barriers.

Traditional software follows predictable patterns, but AI agents introduce uncertainty through adaptive behavior, evolving objectives, and dynamic decision-making. Without proper oversight, these capabilities become enterprise vulnerabilities instead of competitive advantages. The future of work requires AI agents operating alongside humans, but achieving this at scale demands security infrastructure designed for an agentic world.

At General Catalyst, we back founders building entirely new categories that aim to solve critical enterprise challenges. That's why we're co-leading Geordie's seed round as they create the security platform that can enable safe AI agent adoption at enterprise scale.

Securing the Agentic Future

Geordie has built a comprehensive monitoring and governance platform that tracks AI agent activity across enterprise environments. Their system identifies agents regardless of underlying technology, analyzes behavioral patterns, and provides risk assessment in real-time. Through their proprietary intervention system, Beam, enterprises can guide agent actions proactively while preserving operational flow and ensuring policy compliance throughout agent evolution.

A Team Fit for Purpose

Founders Henry Comfort, Hanah-Marie Darley, and Benji Weber created Geordie to allow enterprises to embrace agents safely and securely. From our very first interaction, they exuded energy and enthusiasm. Combine that with deep experience from Darktrace and Snyk, and you get a team poised to tackle this complex challenge. Already, we have seen that energy and technical prowess show itself: from chasing down prospective leads to working with design customers to recruiting an amazing team, the "Geordies" are on a mission to build an iconic company.

Our Conversation with Henry

We recently sat down with Henry to learn more about how Geordie is building the infrastructure that will power enterprise AI adoption at scale. This interview has been edited for length and clarity.

When did you first realize this was the problem you needed to solve?

In early calls with Chief Information Security Officers (CISOs) pioneering agent adoption, two extremes emerged: either they felt unable to enable innovation due to lack of visibility into agent risks, or they admitted they had “put the cart before the horse.” There was no middle ground where risk was well managed, and existing security tools weren’t fit for purpose. It was clear this gap would only grow as enterprises scaled AI agents.

How has your thinking evolved since you started working on this?

We knew enterprises would need controls, but the urgency for autonomous, context-aware remediation has accelerated even faster than expected. Agentic activity quickly scales exponentially once adopted, creating risks that cannot be handled manually. Our focus has therefore accelerated faster from mere visibility and risk intelligence to embedding real-time, adaptive risk mitigations early in the agentic journey.  

From a technical standpoint, what’s novel or fresh about how you’re building?

Geordie understands an agent’s posture and activity deeply and applies dynamic controls that evolve in real time with their specific context, tools, and threats. Unlike static guardrails or blunt interventions that break visibility and auditability, our agent-native approach adapts alongside the agent itself. This lets enterprises safely govern agents without slowing them down.

In early calls with Chief Information Security Officers (CISOs) pioneering agent adoption, two extremes emerged: either they felt unable to enable innovation due to lack of visibility into agent risks, or they admitted they had “put the cart before the horse.” There was no middle ground where risk was well managed, and existing security tools weren’t fit for purpose. It was clear this gap would only grow as enterprises scaled AI agents.

We believe, and have seen throughout our careers, that AI adoption moves at the speed of trust. By giving enterprises the confidence to manage agentic risk in real time, we enable them to innovate boldly while staying safe. Our goal is for Geordie to become a cornerstone of responsible AI adoption worldwide.