A startup called Vijil Inc. said today it has raised $17 million in funding to accelerate the development of its artificial intelligence agent resiliency platform.

The round was led by BrightMind Partners and included participation from Mayfield and Gradient, bringing the company’s total amount raised to $23 million.

Vijil is building the “trust infrastructure” for AI agents, with a framework that helps to guarantee their reliability, security and safety by automating testing and defense mechanisms. Its goal to is to help ensure companies can build and deploy agents quickly and effectively, without having to worry about whether they cause security problems or suffer inaccuracies.

The promise of AI agents is that they act more like people than other forms of AI, performing tasks on someone’s behalf, prioritizing based on the framework they’re supplied with. They’re supposed to adapt to changes in context and seamlessly integrate with third-party tools when required.

But the reality is somewhat different. AI agents hallucinate, they leak sensitive information with abandon, and they create new kinds of vulnerabilities within the otherwise robust networks they operate in. In other words, they make mistakes, and in that respect, indeed they are just like people.

However, the unpredictability of AI agents and what kinds of mistakes they might make creates significant trust issues for enterprises. That ultimately affects how they’re used. Vijil intends to solve this, and it has created quite a few different tools to achieve its goal.

For instance, the startup offers development support with preconfigured agent templates that are customized for different industries, such as human resources and legal departments, which can quickly be adapted based on an organization’s needs. It also supports custom testing, allowing developers to run bespoke tests during development to ensure agents meet their reliability and security standards.

Moving beyond development, Vijil provides tools for continuously monitoring AI agents, as well as runtime defense mechanisms that help to protect against threats such as prompt injection attacks and unsafe outputs. Vijil’s infrastructure provides each AI agent with a “trust score” based on the level of defenses that have been implemented and its testing results, giving users a clear measure of its reliability and safety.

It also carries out regular automated audits of each agent, continuously evaluating its performance and behavior to ensure it doesn’t deviate from established standards. Finally, the company automates compliance with regulations such as the EU’s AI Act and NIST AI Risk Management Framework, generating the documentation they need automatically.

Vijil says organizations will benefit from accelerated deployment because it reduces the time it takes to build and deploy trusted AI agents. Its platform has proven popular, as noted by Gartner Inc., which recently included it in its 2025 Cool Vendors in Agentic AI Trust, Risk and Security Management.

Chief Executive Vin Sharma said the early adoption by customers and the support of its investors validates the company’s vision of an agentic AI trust framework. “We deliver the essential infrastructure layer that enterprises need to trust AI agents in production,” he claimed.

Mayfield Partner Vijay Reddy said Vijil provides enterprises with the tools they need to go beyond experimenting with AI agents and scale them across their entire organization. “Only a small fraction of enterprises are scaling them,” he said. “The biggest barrier is trust, and Vijil is the most comprehensive platform to solve this with continuous learning from observability data.”

Image: SiliconANGLE/Dreamina