Exabeam AI agent security has taken a major step forward. The cybersecurity company has launched new tools to help security teams watch, investigate, and reduce risk from AI agents inside company networks. This is a direct answer to a growing problem: as more AI tools work inside businesses, it is very hard to know if they are behaving safely.
What Did Exabeam Launch?
On July 1, 2026, Exabeam announced updates to its New-Scale Security Operations Platform. The company, based in Broomfield, Colorado, added new features across several parts of its platform. The big news includes 90 total AI-focused detections, support for Anthropic Claude, a new open-source tool called Observra, and smarter investigation tools powered by its AI assistant, Exabeam Nova.
AI agents are software programs that work on their own. They access data, use apps, and take actions without a human doing each step. Because they use valid accounts and approved tools, risky behavior can look normal. That is the problem Exabeam is trying to solve.
Exabeam AI Agent Security: Key New Features
1. Double the AI Detections
Exabeam now has 90 detections focused on AI and agent activity. That is double the number before this release. These detections look for strange behavior, such as unusual tool use, suspicious prompts, and signs that an AI agent may be doing something it should not. One important alert type is called “Denial of Wallet,” which means an AI agent is using too many resources and running up costs, possibly because of an attack.
2. Support for More AI Platforms
The platform can now monitor Anthropic Claude alongside OpenAI ChatGPT, Google Gemini, Microsoft Copilot, and GitHub Copilot. This helps security teams see which AI tools staff are using and spot any new or risky ones.
3. OWASP Coverage for Agentic AI
Exabeam has added coverage based on the OWASP Top 10 for Agentic AI 2026. OWASP is a trusted group that lists the biggest security risks for software and AI. With this update, security teams can see how well their current tools cover those known risks and find any gaps.
4. Observra: A New Open-Source Tool
Exabeam also launched Observra, a free open-source tool. It gives developers and security teams a clean way to track what AI agents are doing. Open-source means anyone can use and improve it at no cost. This is useful for companies that want to build their own AI monitoring without buying extra software.
5. Faster Detection with Exabeam Nova
The Nova Rules Creator lets security engineers write new detection rules using normal language, not complex code. Teams can also convert rules from a popular format called Sigma, which saves time. A new feature called Nova Related Cases, now in early access, helps analysts find connected security events faster by looking for shared details like IP addresses.
Why This Matters for Enterprises
Pete Harteveld, CEO of Exabeam, explained the need clearly. “Organizations are rapidly moving from AI experimentation to autonomous AI agents operating across the enterprise,” he said. “Security teams need visibility not only into human activity, but into how agents behave, interact, and make decisions.”
Andrea Licciardi, Senior Cybersecurity Manager at MAIRE and Founder of CISOs4AI, added a real-world view. “We’re seeing more AI tools and agents show up across the business, and one of the biggest challenges is understanding how they’re actually being used. Behavioral analytics and agent observability give us the context we need to spot unusual activity early and investigate it quickly.”
What Is Behavior Intelligence?
Exabeam calls its approach “Behavior Intelligence.” It is a security model that watches how users and AI agents normally behave over time. When something changes or looks strange, the system raises an alert. This is more powerful than simple rules because it catches new types of attacks that no one has seen before.
Why Pakistani Businesses Should Care
Pakistani companies are quickly adopting AI tools for customer service, finance, and operations. As AI agents become common in local banks, telcos, and startups, the risk of misuse or attack grows too. Exabeam AI agent security tools show the kind of protection modern businesses will need as they grow their use of AI. Understanding these global trends helps local IT and security teams prepare for what is coming.
Quick Summary
Exabeam has expanded its AI security platform with 90 detections, Claude support, OWASP alignment, and the new open-source Observra tool. These updates give security teams better visibility into what AI agents are doing inside their networks and help stop threats before they cause damage.