WhiteMatter just landed $20 million in Series A funding to stop AI agents from going rogue inside enterprise systems.
The 2024 startup, previously called QuilrAI, builds what it calls guardian agents. Each one shadows a company’s existing AI agent and verifies every action against policy before it happens. File access, software use, decision making, even triggered transactions all pass through the guardian first. Outcomes get monitored too.
That oversight layer arrives as enterprises push AI agents deeper into daily operations. More non-technical staff now wield agents capable of executing real tasks. Meanwhile, a wave of security incidents tied to agent misbehavior has sharpened focus on the space. Vidit Arora, WhiteMatter’s 35-year-old founder and CEO, described the stakes bluntly. Agent attacks and compromises, whether external or internal, haven’t even reached “the tip of the iceberg.”
The architecture scales automatically. Every time a company spins up a new AI agent, WhiteMatter generates a matching guardian. Crucially, these guardians don’t consume customer tokens. They run on WhiteMatter’s own hosted model and GPUs. Pricing follows a flat fee based on agents under management, an AI spin on the asset management world’s AUM metric.
WhiteMatter counts Fortune 200 and Fortune 500 firms among its clients, including asset manager Alter Domus. To date, the company reports 2.25 million agents under management and more than 116 billion AI interactions overseen.
The competitive field includes Prompt Security, Lakera, and Protect AI. Blumberg Capital led the round, with Crew Capital and Sprout & Oak participating. Arora plans to direct much of the new capital toward go-to-market hires and sales. The 80-person team should grow by roughly half over the next year, with additional investment flowing into research.
