Druva: Agentic AI Is Already Your Biggest Attack Surface

Name: Druva: Agentic AI Is Already Your Biggest Attack Surface
Uploaded: 2026-09-01T08:25:03-04:00
Duration: 1 min 13 s
Description: Druva CEO Jaspreet Singh breaks down why agentic AI has quietly become the largest attack surface in the enterprise: when an AI agent has access to your email, it has access to your two-factor authentication — and from there, every application in your ...

Druva

09/01/2026

0 (0%)

Report Like Favorite

TL;DR

Agentic AI systems that act autonomously on behalf of users already have broad enterprise access — this is a present-day risk, not a future concern.
Granting an AI agent access to email effectively grants it access to two-factor authentication, which is the gateway to every enterprise application.
Even in well-secured environments, the sheer size of the attack surface created by AI agent permissions represents a significant risk management challenge that grows as AI adoption increases.

Summary

In this short executive clip, Druva CEO Jaspreet Singh delivers a pointed warning about agentic AI and the enterprise attack surface it creates today — not in some hypothetical future. Singh explains that as organizations integrate AI agents capable of acting autonomously on behalf of users, they are inadvertently granting those agents access to email inboxes. Because email is the delivery channel for two-factor authentication codes, that single permission effectively unlocks every application in the enterprise environment. The attack surface this creates is not incremental — it is exponential. Singh acknowledges that well-secured, tightly managed applications reduce the likelihood of exploitation, but frames the issue through a risk mitigation lens: the surface area is already extremely large, and as enterprise adoption of autonomous AI accelerates, the risk factor amplifies rather than stabilizes. The message is a clear call for security leaders to reassess how AI agent permissions are scoped and governed before autonomous access becomes an unmanageable liability.

Chapters

0:00 - AI Access Is Already Here
0:09 - How Agentic AI Works
0:26 - Email, 2FA, and Full Enterprise Exposure
0:49 - Risk Amplification as AI Autonomy Grows

Key Quotes

0:00 "I think AI already has access to all the applications. It's not a future set statement. It's a current statement."
0:26 "If you've already given an email access, you're practically given the AI bot access to your two-factor authentication, which is the genesis of an access to any application in enterprise."
0:49 "Security is all about risk mitigation. From a risk mitigation perspective, the surface area is extremely large."

FAQ

Why is email access so dangerous to grant an AI agent?

Email is the delivery channel for two-factor authentication codes. If an AI agent can read your email, it can intercept 2FA tokens, which effectively gives it authenticated access to any enterprise application that relies on email-based verification.

Does this mean enterprises should stop using agentic AI?

Singh does not advocate for halting AI adoption. Instead, he emphasizes risk mitigation — acknowledging that tightly managed applications reduce exploitation likelihood, but stressing that the attack surface is already very large and grows as AI autonomy increases.

Categories:

Tags:

Show more Show less

TL;DR

Summary

Chapters

Key Quotes

FAQ

Bridging the SaaS Protection Gap: Preventing Data Loss and AI Missteps

The Automation Escalation: Discovering the AI-Driven Underground Revolution

Invisible Data: Understanding What You Can't Safeguard