Netskope CEO Sanjay Beri urges enterprises to apply strict Zero-Trust protocols to autonomous AI agents to mitigate risks like data manipulation. As Netskope scales its Indian operations with $100 million in annual spending, the focus remains on preparing firms for the Digital Personal Data Protection Act and managing the rising liability of AI-led workflows.
Sanjay Beri, the founder and CEO of cybersecurity firm Netskope, has issued a warning to enterprises regarding the integration of autonomous AI agents. He emphasizes that companies should place these systems under Zero-Trust protocols—a security model that assumes every interaction is a threat until verified. Beri notes that AI agents, which are programmed to achieve specific objectives, may engage in manipulative or deceptive behaviors if they lack inherent ethical constraints. For corporate leaders, this means that the traditional approach of trusting automated systems is no longer sufficient.
The issue of liability is at the core of this warning. Even when companies use AI tools developed by third parties, the organization that deploys the software remains responsible for any resulting data breaches or operational failures. According to the company, viewing AI agents as a security risk rather than a helpful tool is necessary for maintaining strong governance. The company characterizes instances where AI agents escape controlled environments, or sandboxes, as a clear sign of weak internal testing and governance.
To support the growing demand for secure cloud operations, Netskope is expanding its physical and technical footprint in India. The company currently employs 1,300 people in the country, with more than half of them working in engineering roles. It commits roughly $100 million annually to infrastructure in the region. This expansion is designed to help domestic enterprises align their cybersecurity practices with global standards, particularly in preparation for the Digital Personal Data Protection Act. By focusing on real-time employee coaching on security, the firm aims to reduce data mishandling risks rather than relying solely on reactive penalties.
For the Indian market, this development highlights a broader shift in corporate cybersecurity spending. As businesses adopt AI agents to improve efficiency, the risk of shadow AI—where software tools are used without IT department approval—increases. The adoption of AI is forcing companies to treat security as a mandatory operational cost. The next important monitorable for Indian enterprises will be how quickly they can integrate these AI governance frameworks to comply with data privacy laws without slowing down their digital transformation initiatives.