Zscaler Launches Industry-First Zero Trust Security for Agentic AI
LAS VEGAS – Zscaler Inc. on Tuesday updated its flagship Zero Trust Exchange platform to secure how artificial intelligence (AI) agents connect, access data, and operate across devices.
As organizations rapidly adopt agents, traditional security frameworks built around predictable human identities are struggling to keep pace. Autonomous agents can operate independently, spawn sub-agents, and create temporary identities, introducing critical visibility and governance gaps.
To address these vulnerabilities, Zscaler introduced two major solutions at its annual developers conference here: Zscaler AI Broker, which regulates agent-to-agent communications and enforces fine-grained access policies via a centralized Agent Registry, and Zscaler Endpoint AI Security, which detects hidden threats within browsers, extensions, and local AI tools that legacy endpoint software often misses.
Anchoring the launch is Zscaler AI Access Graph. Undergirded by the company’s recent acquisition of Symmetry Systems, the tool maps the intricate relationships between users, AI agents, applications, and data sources. This allows organizations to track data lineage in real-time and aggressively reduce unnecessary access risks.
“CIOs and CISOs want visibility badly to discover, secure, and harden their operations,” Swamy Kocherlakota, executive vice president of agentic cloud/AI for Zscaler, said in an interview.
Additionally, Zscaler announced substantial upgrades to its AI Protect portfolio, which launched earlier this year. The enhancements span three core areas. AI Asset Management expands visibility into embedded AI within SaaS traffic, public clouds, and endpoint activity. Secure Access to AI enhances guardrails for over 250 generative AI apps, including compliance API support for OpenAI and Anthropic. Secure AI Infrastructure includes new AI red teaming capabilities and prompt-hardening services to secure applications throughout their development lifecycle.
“Agents are often the weakest link,” Deepen Desai, Zscaler’s chief security officer, said in an interview. “It doesn’t realize it’s been compromised and become a malicious actor. The cat is out of the bag.”
Industry leaders emphasized that securing these autonomous ecosystems is the next frontier of cybersecurity.
“Traditional security was never designed for millions of autonomous agents that act and reach sensitive data at machine speed,” Zscaler CEO Jay Chaudhry said. “We pioneered Zero Trust Exchange to secure users, branches, and cloud workloads, and now we are innovating to extend that security to AI agents. Enterprises are no longer held back from rolling out agents everywhere.”
John Israel, global chief information security officer at KPMG, echoed the sentiment during the announcement, noting that managing modern data security requires scaling visibility rather than just building higher walls.
“As businesses scale their use of AI agents to optimize operations, having a unified, zero-trust framework to trace data lineage and govern agent-to-agent interactions is paramount to maintaining trust, compliance, and competitive advantage,” Israel said.
