As AI becomes part of everyday business operations, security teams are running into a clear problem. Access and security models designed for users and applications do not easily cover AI agents, browser-based AI tools, and custom models. Data moves through more paths, visibility breaks down, and policies are enforced unevenly.
Cloudbrink’s latest platform updates focus on bringing AI traffic into the same secure connectivity and ZTNA framework already used to protect the hybrid workforce.AI creates new security gaps
AI systems introduce new ways for sensitive data to move in and out of the enterprise. Agents can run locally, in the cloud, or inside applications. Online AI services often sit outside traditional enterprise controls. Many organizations respond by adding more tools, which increases complexity without solving the root problem.
Cloudbrink CEO Prakash Mana says consolidation is about more than reducing tools. He told ChannelE2E, “Cloudbrink provides a unified, in-depth security control and deep visibility not only to the new threats posed by AI adoption across AI Agents, AI Tools, Applications or any other form factor, but also end-to-end insights across the stack - from the network layer, user identity, and user access to insights into AI protocols, prompts, and user context,” Mana said.
That level of visibility is difficult to achieve when security controls are spread across disconnected products.
Bringing AI into the ZTNA model
Cloudbrink is extending its ZTNA platform, so AI agents and AI services are governed by the same access and connectivity controls as users and applications. Instead of treating AI as a special case, it becomes part of the access layer itself.
According to Mana, this approach avoids the limitations of stitched-together security stacks. “Cloudbrink’s ability to control and provide insight across all of these users at a level of granularity that includes not only user, device, browser, but also at the micro agent level within applications is not possible by stitching together multiple security tools,” he said. Secure browsers and network proxies tend to see only part of the picture, which can limit controls or hurt productivity.
Real-time policy enforcement for agentic AI
Traditional DLP and proxy-based tools often lack the context needed to manage AI traffic effectively. Cloudbrink’s platform enforces policy in real time using a broader view that includes user, device, agent, network, location, environment, and AI-specific context.
“Comprehensive user, device, agent, network, location, environment, enterprise, and AI context are extremely important as we enable fine-grained security controls and provide insights,” Mana said. “Traditional DLP, proxy-based controls, or segment-specific solutions miss one or more of these contexts that prevent comprehensive AI controls, visibility, and insights.”
This allows security teams to control how AI systems access data as activity happens, rather than reacting after data has already moved.
One platform, shared visibility
Cloudbrink also brings users, applications, AI agents, online AI services, and traffic into a single management console. This gives security teams clearer insight into how data is accessed and used, without having to correlate information across multiple tools.
As Mana put it, “The need for a consolidated platform is not just about simplicity but that of effectiveness, overall security, and productivity while enabling AI in the enterprise.”
What this means for MSPs and partners
For MSPs and partners already delivering ZTNA, Cloudbrink positions AI security as part of the same managed service, not a separate layer to deploy and operate.
“From a solution delivery perspective, the integrated AI abilities can be offered seamlessly on the exact same platform they have been delivering to customers,” Mana said. “No additional complexity to delivering, managing the solution.”
While the service model remains familiar, the scope expands. AI security becomes a core capability, opening new opportunities where customers are looking for a single platform that can secure users, applications, and AI agents together.
AI is changing how data moves through the enterprise, and security models need to adapt. By extending ZTNA to cover AI agents and services, Cloudbrink is pulling AI back into the core access and connectivity framework. For organizations and partners supporting hybrid work, this approach reduces blind spots and makes AI security part of the foundation rather than an add-on.
