90 Days With Frontier AI: What We Learned

When Anthropic and OpenAI gave Zscaler access to their Frontier AI models, we didn't just study the threat. We got to work.

Over the last 90 days, our team ran assessments across dozens of leading enterprises, operating the same way an adversary would: scanning attack surfaces, chaining vulnerabilities, and stress-testing architectures that have protected organizations for years.

Join Zscaler security leaders as we share what 90 days of real-world assessments actually taught us — and what it means for how you build your defenses from here.

You’ll learn:

• What the data shows — patterns across dozens of real enterprise assessments that reveal where organizations are exposed and where they're holding up
• What surprised us — the findings that cut against conventional wisdom and changed how we think about AI-era risk
• What the prepared organizations are doing differently — the architectural decisions that separate resilient organizations from exposed ones