Artificial Intelligence & Machine Learning
,
Data Security
,
Events
Netskope’s Sanjay Beri on Data Risk, Agent Visibility and Enabling AI Safely
Enterprises aren’t securing artificial intelligence as much as catching up to it. Security teams are reacting to AI adoption the way they once reacted to cloud, said Sanjay Beri, co-founder and CEO at Netskope.
See Also: Taming the Rise of Shadow AI Agents
Data loss is the primary risk. Employees routinely submit sensitive information to public large language models because the productivity benefit is immediate and guardrails are not yet in place.
“Data classification by itself is useless. What you need is context. How is this data moving? Where is it going? Once you have that context, you can put guardrails that let people unleash your data in the right place, yet stop it from being used in the wrong way,” Beri said.
In this video interview with Information Security Media Group at RSAC Conference 2026, Beri also discussed:
- How real-time monitoring of prompts and outputs reduces data exposure;
- Why 90% of AI usage is business unit- or user-led, not IT-controlled;
- Why geopolitical volatility and nation-state cyber activity have accelerated demand for data sovereignty.
Beri has more than two decades of innovation and success in the cloud, networking and security industries. Prior to starting Netskope in 2012, Beri was the vice president and general manager of Juniper Networks’ secure access business unit, driving it to market leadership throughout his tenure. He also co-founded one of the world’s first data center encryption companies, Ingrian Networks.

