Manifold Security appoints Joe Sullivan to its board

Manifold Security has appointed Joe Sullivan to its board of directors. Sullivan is a former chief security officer at Facebook, Uber and Cloudflare.

The appointment adds a veteran cybersecurity executive to the governance of a company focused on risks linked to AI agents in the workplace.

Sullivan brings more than three decades of experience across the public and private sectors. Before his roles in Silicon Valley, he served as a federal prosecutor at the US Department of Justice and helped establish the Computer Hacking and Intellectual Property Unit in the Northern District of California.

His arrival comes as Manifold expands its work on monitoring and controlling AI agents used across employee devices, browsers, desktop applications and cloud environments. He was also an early backer and is among the investors behind the USD $15 million the company has raised so far.

AI security

Manifold is positioning itself in a market drawing greater attention as businesses test software agents that can carry out tasks with growing autonomy. Security concerns have intensified as major AI developers push these products beyond closed demonstrations and into broader use.

The company cited a Deloitte estimate that 74% of companies expect to use AI agents by 2027. Against that backdrop, businesses need visibility into which agents are active across their organisations, rules governing what those agents are allowed to do, and the ability to detect and respond to threats as they emerge.

Manifold says it has already worked with multiple Fortune 500 companies to secure their use of AI agents. It is now extending that coverage across endpoints, browsers and the cloud, where such systems may operate within an organisation.

Sullivan outlined his view of the risks posed by AI agents.

"AI agents are a fundamentally different security challenge because they're built on probabilistic software. We can send an agent in to do a job and it will do the right thing most of the time, but occasionally it will go off the reservation. Manifold Security recognized early on that this wasn't a future problem. It was a problem enterprises would be facing as soon as they started deploying agents at scale. They were thinking about these risks ahead of the rest of us, and have built a solution at exactly the moment organizations are starting to confront them. That's why I'm excited to join the board and help the team as this market develops," said Joe Sullivan, board member, Manifold Security.

Founders' background

The appointment also highlights the backgrounds of Manifold's founders, Neal Swaelens and Oleksandr Yaremchuk. Before starting the company, they launched Laiyer AI, where they created LLM Guard, an open-source firewall designed for large language model systems.

Manifold's leadership said Sullivan's mix of legal, investigative and corporate experience makes him a strong fit as companies weigh how to govern increasingly autonomous software within their operations.

"Joe has lived through virtually every phase of the modern security industry, from investigating cybercrime as a federal prosecutor to building security programs at some of the world's largest technology companies. He understands both the technical reality of what agents are doing and the organisational consequences when security teams cannot see or control those actions. We couldn't think of a better person to have alongside us as we build the security layer for the agentic workforce," said Neal Swaelens, co-founder and chief executive officer, Manifold Security.

Yaremchuk linked that challenge to the broader advance of AI systems that can take on more work with less direct human input. He said the pace of adoption is increasing pressure on security teams to track and govern what those systems are doing inside companies.

"We're reaching an inflection point in agentic AI capabilities, with the potential for enormous gains alongside a new class of security and safety challenges. With the launch of GPT-6 Astra, a new standard has been set for how much work can be offloaded to an AI agent, which will only increase as AI advances. Organisations are moving quickly to harness these extraordinary capabilities, making it increasingly important that security keeps pace. We built Manifold around this exact challenge, giving companies visibility and control over what their agents are actually doing as they operate across the enterprise. Joe has long shared our vision for securing the agentic workforce, and we're pleased to have his experience and perspective alongside us as we build what comes next," said Yaremchuk.