Vijil Raises $17 Million to Build the Trust Infrastructure for Enterprise AI Agents

Vijil, a San Francisco–based startup building trust and safety infrastructure for enterprise AI agents, has raised $17 million in funding led by BrightMind Partners, with participation from Mayfield and Gradient. The new round brings the company’s total funding to $23 million and will be used to accelerate adoption of its platform, which helps organizations build, test and operate AI agents that are reliable, resilient and safe for production use.

The announcement comes at a moment when enterprises across industries are experimenting with autonomous and semi-autonomous AI agents, yet few are able to deploy them at scale due to concerns about security, reliability and governance. Vijil positions itself as the technology layer that addresses these concerns directly, providing what it describes as a “trust infrastructure” for AI agents.

Addressing the Growing Trust Gap in AI Agent Deployment

While AI agents promise productivity gains across hiring, operations, customer support and workflow automation, enterprises remain hesitant to put them into real-world environments. Large language models (LLMs) can be misled, manipulated or produce harmful outputs; they often lack the guardrails needed for enterprise-grade deployment. According to Vijil, most organizations exploring AI agents today face a similar set of challenges:

• Limited expertise in securing and governing agentic workflows
• High costs associated with manual testing and compliance
• Inability to reliably detect unsafe behaviors or failures
• Lack of continuous learning from production telemetry
• Long deployment cycles due to risk-related hurdles

Vijil’s platform aims to solve these issues end-to-end, from development to production.

A Platform Designed to Test, Defend and Continuously Harden AI Agents

Vijil offers a modular stack that enterprises can use to build, test, deploy and continuously improve their AI agents. Its platform includes two cornerstone products:

Vijil Evaluate: An automated testing framework for LLM applications

Vijil Evaluate provides quality assurance for AI agents, automating stress tests, behavioral tests and compliance checks. The framework aims to reduce the manual effort teams spend ensuring reliability and security before deployment. According to the company, Vijil Evaluate shortens “time-to-trust” for AI agents by enabling developers to run automated evaluations that would traditionally require weeks of manual validation.

Vijil Dome: Real-time runtime defense for agent safety

Vijil Dome monitors AI agents during live operation, detecting and blocking out-of-policy or unsafe inputs and outputs. Dome acts as a guardrail layer, preventing harmful behaviors from propagating into business workflows. This real-time defensive posture reflects a growing recognition across the industry that AI agents are not static, they require continuous operational oversight.

A Seasoned Team With Deep AI Infrastructure Experience

Vijil was founded in 2023 by senior technologists who previously built large-scale AI and deep learning infrastructure at AWS (Amazon SageMaker) and Splunk SEAL. The team is advised by academic researchers and AI security experts known for pioneering work in deep learning and LLM safety. The company emphasizes a commitment to open innovation, citing collaborations with academic institutions, nonprofit research organizations and compatibility with open foundation model ecosystems. Its investors: BrightMind Partners, Mayfield and Gradient, all bring strong expertise across AI infrastructure and enterprise software.

“Vijil has assembled a seasoned team with deep experience in AI infrastructure at AWS,” said Stephen Ward, general partner at BrightMind. “What sets Vijil apart is the ability to continuously harden AI agents through reinforcement learning from production telemetry.”

Enterprise Adoption and Early Customer Impact

Vijil reports growing adoption among enterprise AI teams, including companies deploying AI agents in sensitive workflows where trust verification is essential. One customer, SmartRecruiters, said Vijil’s platform helped compress deployment timelines significantly.

“Our enterprise customers demand trust verification before deploying AI in hiring workflows,” said Michal Nowak, SVP of Engineering at SmartRecruiters. “Vijil helps us ship AI agents in six weeks instead of six months while dramatically lowering compliance costs.”

While these figures reflect company-reported results, they underscore the central challenge many enterprises face: ensuring AI agents behave safely in real-world environments.

Recognition as a Gartner Cool Vendor

Vijil was recently recognized as a Gartner® Cool Vendor™ in the 2025 Agentic AI Trust, Risk and Security Management (TRiSM) category. The recognition signals growing interest in the emerging discipline of AI agent governance, a domain that analysts expect to become critical as companies transition from AI prototypes to production-grade autonomous systems.

The funding announcement and analyst recognition together highlight Vijil’s positioning within a larger shift in the enterprise AI landscape. Trust is becoming as important as model performance.

Building the Missing Infrastructure Layer for Safe AI Agents

The company argues that existing AI safety tools, many of which focus on individual components like monitoring or red-teaming are insufficient for the dynamic, interactive behaviors of AI agents. Vijil’s platform is built around continuous reinforcement learning from operational telemetry, allowing agents to improve resilience over time.

“Most enterprises are experimenting with AI agents but only a small fraction are scaling them,” said Vijay Reddy, partner at Mayfield. Vijil’s founder and CEO, Vin Sharma, described the company’s mission clearly: “Vijil delivers the essential infrastructure layer that enterprises need to trust AI agents in production.”

With the fresh funding, Vijil plans to expand deployment efforts, grow its engineering team and deepen integrations with ecosystem partners across the AI industry.

Vijil arrives at a critical moment for enterprise AI: organizations are eager to deploy agents but lack the infrastructure to secure, monitor and validate them. The company’s full-stack approach, testing, runtime defense and continuous reinforcement learning, positions it well within the rapidly emerging TRiSM category. However, the ultimate test will be scale. As AI agents become more complex and handle more autonomous decision-making, real-time defenses and evaluation frameworks must adapt just as quickly.

If Vijil can deliver consistent trust-building results across diverse enterprise environments, it could become one of the defining companies in AI agent governance.