Everseen Joins Industry Leaders Behind NVIDIA's Open Agent Safety Platform

On September 28, 2026, NVIDIA launched the Open Agent Safety Platform, an open reference design for monitoring and controlling autonomous AI agents. Everseen is among the companies across the AI ecosystem supporting the initiative.

This development comes at a crucial time as AI agents grow in sophistication and in what they are trusted to do. They now call tools, access data, and carry out complex multistep tasks autonomously. Several frontier AI labs have recently reported agents acting beyond their designed environments.

Open Agent Safety Platform creates necessary AI safeguards

NVIDIA's answer, the Open Agent Safety Platform, combines two layers of protection. NVIDIA OpenShell is runtime software that creates and enforces boundaries for AI agents. NVIDIA Sentry adds an independent layer by monitoring agents continuously and stopping any that attempt to move outside their pre-defined boundaries. These safeguards extend across the agents' operations, whether they are being tested in a controlled environment or running live in production.

We believe this is a crucial development because agents cannot be expected to fully govern their own behavior; safety controls need to exist outside the agent, beyond its reach. Independent oversight like this makes it possible to safely develop agents and unlock what they can do. It also gives organizations a path from limited pilots to adoption at enterprise scale, in retail, and well beyond.

Everseen are long-standing NVIDIA partners and our Vision AI runs on NVIDIA accelerated computing. NVIDIA's clear commitment to AI safety further reinforces our dedication to that partnership.

Proud to support an industry-wide push for safer agents

Responsible AI has always been core to how we build at Everseen, and we are glad to stand with the industry leaders prioritizing AI safety. Anthropic has collaborated with NVIDIA to add further layers of security and control, and organizations from Hugging Face to CrowdStrike are among more than 100 working with the platform's technologies. 

As Everseen CEO Joe White said in response to the announcement, "Nobody scales what they can't control, so responsible AI measures are what will confidently move agents from experiments to doing the extraordinary work they're capable of. Proud to be a partner and to share the belief that trust and innovation go hand in hand."

This philosophy shapes our own platform, which is privacy-by-design and built to be GDPR and EU AI Act compliant. We look forward to working with NVIDIA toward a future where AI is both trusted and capable.

Read more in the Open Agent Safety Platform technical overview

‍