Certiv Raises $4.2M in Pre-Seed Funding to Secure AI Agent Activity on Endpoints

Seattle has always had a certain rhythm to it. Builders in hoodies, engineers with quiet grins, the sound of something serious being assembled behind closed laptops. Then every so often a signal cuts through the fog. This week that signal is Certiv, a 9 person startup that stepped out of stealth with $4.2M in pre seed funding and a very timely idea about what happens when AI agents start roaming around your company computers like interns who never sleep.

Credit where it is due. Jason Needham, Co-founder & CEO, Paul Allen, Co-founder & CTO, and Daniel Morris, Co-founder & Chief AI Officer (CAIO), just turned a sharp observation into a funded mission. The round was led by Aviso Ventures, the firm founded by Signal Sciences creator Andrew Peterson, with Founders Co op, Fortson, and additional investors joining the table. When investors with deep security instincts start writing checks this early, it usually means they see the same storm clouds forming on the horizon.

Here is the problem hiding in plain sight. AI agents can write code, read files, and move through systems using employee credentials. Useful, powerful, and slightly terrifying if nobody is watching the door. Traditional security tools were not designed for software that can think, act, and improvise inside a laptop. That gap between capability and control is exactly where Certiv decided to set up shop.

Certiv’s platform runs directly on employee machines across Windows, Mac, and Linux. Think of it as a digital gatekeeper with a very strict guest list. Every action an AI agent attempts gets evaluated against company policy before it reaches sensitive systems. If something crosses the line, it gets blocked. Not debated. Not logged for later. Stopped. The platform also detects unauthorized agents, tracks the chain of reasoning behind what an agent is trying to accomplish, and gives organizations visibility into a new category of digital worker that most companies barely understand yet.

The company calls this runtime assurance for AI agents, which sounds technical until you realize what it really means. It means someone is finally asking the question that matters. If an AI agent is acting with your employee’s credentials, who is actually in charge of the keyboard?

Founded in June 2025 and operating as a remote first team based out of Seattle, with time spent inside the Foundations startup community on Capitol Hill, Certiv is already running multiple enterprise pilot deployments. The new capital will expand the engineering team and push those early deployments further into the wild where real enterprise systems live and breathe.

The bigger takeaway for founders is simple. Every new wave of technology creates its own security problem a few months later. The companies that notice the gap early are usually the ones investors chase first. AI agents are about to become coworkers in a lot of organizations, and coworkers tend to do unexpected things when nobody is watching.