Varonis Adds Browser, Endpoint and AI Gateway Protection

The new capabilities extend Varonis Atlas coverage to where employees use AI and where AI connects to models, tools, and data.

Updated on Nov. 9, 2026 in Cybersecurity

Varonis Adds Browser, Endpoint and AI Gateway Protection

Live Poll

Do you trust your employer's AI tools to keep your work data safe?

Varonis Systems announced three new capabilities for Varonis Atlas: Browser Protection, Endpoint Protection, and an AI Gateway, all available now. The company also expanded Atlas Agent Intent with a new Agent Manifest that defines what each AI agent is for.

Why it matters

As employees adopt AI tools daily, many of those tools operate outside sanctioned systems and can reach sensitive data without oversight. The additions aim to give security teams visibility and controls across browsers, laptops, and the traffic between AI systems, models, and tools.

Browser Protection operates through a lightweight browser extension, while Endpoint Protection runs on the Varonis endpoint sensor or integrates with CrowdStrike, SentinelOne, and Microsoft Defender. The AI Gateway routes requests across model providers, caps token spend, and enforces guardrails on prompts, responses, and agent tool calls, but the announcement offers no independent performance benchmarks.

The players

Varonis Systems

Varonis Systems is a cybersecurity company focused on data security and threat detection, and it markets the Atlas platform for protecting enterprise data.

CrowdStrike

CrowdStrike is a cybersecurity firm known for its endpoint detection and response platform, one of the third-party sensors Varonis Endpoint Protection can integrate with.

SentinelOne

SentinelOne is a cybersecurity company that provides endpoint protection software, another third-party integration available within Varonis Endpoint Protection.

Microsoft Defender

Microsoft Defender is Microsoft's endpoint security product suite, which Varonis Endpoint Protection can integrate with.

The details

Browser Protection finds AI use in the browser, including personal accounts and unsanctioned apps, and blocks unsanctioned AI sites while applying data-aware guardrails. Endpoint Protection locates AI apps, coding agents such as Claude Code, local model files, and MCP servers on laptops and desktops, and the AI Gateway places an LLM and MCP gateway in front of AI traffic. Atlas also uses intent-based access control to step in when an agent drifts from the purpose defined in its manifest.

November 9, 2026: Varonis announced the Atlas browser, endpoint, and AI Gateway capabilities.

The Tech Race

The additions extend security coverage to the Model Context Protocol (MCP), the open standard that now links AI models to external tools and data, and Varonis is positioning Atlas against a field of vendors racing to govern agentic AI. The move reflects a broader industry shift from securing networks alone toward governing the AI agents and connections that operate inside them.

Employees using AI assistants in browsers or coding tools may find their sessions monitored or blocked when they reach unsanctioned services, and personal accounts can be flagged under the new controls. IT teams gain the ability to set spending caps and guardrails on AI usage, which may change which tools staff can access at work.

The takeaway

Organizations adopting AI tools should inventory where those tools run, including personal accounts and local model files, before relying on any single control. Defining what each AI agent is meant to do gives security teams a clearer baseline for spotting when behavior has drifted.