Zscaler teams up with OpenAI to secure enterprise AI at scale
Cybersecurity firm Zscaler expanded its AI-Guardian project on June 9, 2026, adding OpenAI and AWS as partners. The goal is to help companies adopt AI tools safely without risking data breaches. OpenAI's GPT models are used to automatically find and fix security vulnerabilities.
Zscaler is a security company known for its zero-trust approach — instead of trusting everyone inside a company network, it checks every user and device before granting access. The new expansion of AI-Guardian brings OpenAI's models directly into Zscaler's security workflows to detect threats faster and prioritize which vulnerabilities to fix first.
On the operations side, Zscaler is deploying AI agents (automated programs) alongside human security analysts in what it calls Agentic SecOps. This combination keeps the rate of correctly identified real threats at 99.6%. Additionally, through a program called Trusted Access for Cyber, companies can let employees use external AI tools like ChatGPT while Zscaler ensures sensitive company data does not leak out in the process.
Key points
- Zscaler announced AI-Guardian expansion with OpenAI and AWS on June 9, 2026
- OpenAI GPT models automatically detect security vulnerabilities and rank which to fix first
- AI agents work alongside human analysts in Agentic SecOps, achieving 99.6% true-positive threat detection
- Trusted Access for Cyber lets employees use external AI tools safely without data leaks
- Combines zero-trust security with AI to strengthen cloud-based enterprise protection
Quick term guide
- cybersecurity
- The work of protecting computers, software, and online systems from attacks.
- AI tools
- Software that can help create text, code, images, or other work.
- vulnerabilities
- Weaknesses in a computer system that hackers can exploit.
- zero-trust
- A security approach where no user or device is automatically trusted — everyone must verify their identity every time they access a system.
- AI agents
- AI agents are AI tools that can carry out steps toward a goal, not just answer once.
- Agentic SecOps
- Security operations where AI agents (automated programs) work alongside human analysts to detect and investigate cyberattacks.
- Trusted Access for Cyber
- A verified pathway that lets company employees use external AI services while preventing sensitive business data from being exposed.
- enterprise
- A large business or company, which usually buys special software plans for better security and privacy guarantees.