Runlayer raises $30 million to oversee enterprise AI agents

Runlayer has raised $30 million in Series A funding as companies look for ways to control the growing use of AI agents inside the workplace, Fortune reported. The round gives the young startup more capital to build software that monitors what those agents can access, what they do and what they cost.

Felicis led the round, with Khosla Ventures also taking part, according to Fortune. The financing brings Runlayer’s total funding to $42 million, and Fortune reported that Felicis moved early to secure the deal.

Fortune reported that Vinod Khosla wanted “to buy every available dollar of the round” after learning Runlayer was raising money. The company had come out of stealth seven months earlier with an $11 million seed round backed by the same two investors.

A control layer for workplace AI

Runlayer is aimed at a problem that has become more urgent as businesses put AI agents into daily operations. Fortune described the core issue as a lack of dependable tools for companies to track agent behavior, limit data access and prevent costly or unsafe activity.

One Runlayer customer found last year that an AI agent had consumed its full annual AI-compute budget over a single weekend while stuck in a loop, according to Fortune. The incident points to the kind of oversight gap Runlayer is trying to close.

The product lets employees use AI tools such as OpenAI’s ChatGPT, Anthropic’s Claude, Salesforce’s Agentforce and custom agents through company-approved connections, Fortune reported. Security and IT teams can view agent activity, see which data was accessed and track spending from one dashboard.

Runlayer also flags unauthorized AI tools used by employees, according to Fortune. Reports cited by Fortune suggest that such “shadow AI” use could affect as many as 78% of enterprise AI users.

Chief Executive Andrew Berman told Fortune that workers will hand off tasks to “swarms of agents” as a normal part of work. Berman founded Runlayer about 10 and a half months ago, Fortune reported.

Founder background and early customers

Berman previously co-founded Nanit, the AI baby monitor company that had estimated revenue above $100 million as of 2024, according to Fortune. He also co-founded Vowel, an AI video conferencing company that raised $17.8 million before Zapier bought it in 2023.

After the Vowel deal, Berman served as Zapier’s director of AI and worked with teams at OpenAI and Anthropic, Fortune reported. He told Fortune he signed Runlayer’s term sheet on the day his first child was born, saying the company was seeing strong enterprise demand and repeated versions of the same customer problem.

Runlayer has signed more than 12 unicorn customers, according to Fortune and the company’s website. Fortune named Instacart, Gusto, Opendoor, dbt Labs and AngelList among its customers, and reported that half of Gusto uses the product daily.

The company has also moved into financial services, Fortune reported. A Fortune 500 bank is using Runlayer to monitor AI activity for more than 100,000 employees across 200,000 devices, according to the report.

A crowded market

Gartner projects that 40% of enterprise applications will include AI agents by the end of 2026, up from less than 5% in 2025, according to figures cited by Fortune. Forecasts cited by Fortune also put agentic AI spending on track to reach $201.9 billion this year.

Competition is building around agent governance. Fortune reported that Wiz, Palo Alto Networks and Okta are developing related capabilities, while the agentic AI security market is valued at $55 billion in 2026 and projected to reach $888 billion by 2035.

Felicis general partner Jake Storm, who led both Runlayer rounds, told Fortune that governance can help companies adopt agents rather than act only as a cost. He said a neutral control layer is needed because no single platform can own the way agents work across providers.

Berman declined to share revenue figures with Fortune. He said Runlayer is nine months ahead of the product plan shown to seed investors and has expanded beyond Model Context Protocol-focused tooling into a broader interoperability layer for enterprise agents, plugins and skills.

This story draws on original reporting from Fortune.