As organizations rush to adopt generative AI tools, the humble web browser has quietly become one of the most critical, and vulnerable, points in the enterprise security stack. Dhawal Sharma, executive vice president for product strategy at Zscaler, discusses how browser-based workflows, AI tools and unmanaged devices are reshaping the threat landscape and forcing security teams to rethink traditional controls.
For many organizations, the browser has become the default interface for everything from SaaS applications to developer tools and generative AI assistants. That shift creates new attack vectors that conventional security tools, such as firewalls, endpoint detection platforms or signature-based defenses, often struggle to detect. Threat actors are increasingly targeting browser behaviors through techniques like malicious extensions, identity-based phishing and “last-mile reassembly,” where seemingly benign components combine on a user’s device to form a malicious payload.
The growing use of AI is adding another layer of complexity. Employees are rapidly adopting AI-powered tools, extensions and embedded assistants across SaaS platforms, often without direct oversight from security teams. This phenomenon, sometimes called “Shadow AI,” expands the attack surface and raises the risk of sensitive data exposure through prompts, uploads or automated workflows.
Sharma notes that many organizations are still catching up with the pace of AI adoption. Security teams are moving beyond traditional signature- and pattern-based detection toward intent-based controls designed to identify risky behavior rather than specific attack signatures. That shift reflects the reality that AI-driven attacks, and the tools employees use, are evolving too quickly for static defenses to keep up.
At the same time, regulatory frameworks and compliance requirements are beginning to address AI-related risks, pushing organizations toward continuous monitoring and risk assessment across their AI ecosystems. Ultimately, Sharma argues, securing modern enterprise environments will require treating the browser as a central control point for data, identity and application access, especially as AI-driven workflows continue to expand.