What does it really take to secure AI as it becomes part of the fabric of how we work? I had the chance to sit down with Sunil Agrawal, Chief Information Security Officer at Glean — and the company’s first-ever CISO — to dive deep into this question.

Sunil’s perspective is clear: securing AI isn’t about slapping on controls after deployment. It’s about building secure AI by design — from prompt injection defense to real-time content protection and fine-grained governance.

In our conversation, we covered a lot of ground:

Why traditional security frameworks fall short for AI agents

How Glean is preventing real-world attacks like jailbreaks and prompt injections

Why content awareness is key to managing sensitive data exposure

What true visibility and control over agents looks like in the enterprise

Why trust — not just compliance — is the new security currency

And how their new partnership with Palo Alto Networks is raising the bar on threat detection and posture management

With over 100 million actions now executed by Glean’s AI agents each year, it’s no longer a hypothetical risk landscape — it’s a live one.

If you're responsible for security, IT, or AI strategy at your org, I highly recommend giving this interview a watch. It’ll reshape how you think about governance, risk, and the future of work.

Let me know what resonated with you — and if you're thinking differently about securing your AI initiatives after this.

#data #ai #agents #GleanGo #theravitshow