Zscaler builds out AI security services opportunity for MSPs

Cybersecurity vendor Zscaler expanded its Zero Trust Exchange platform by bringing in new endpoint AI security capabilities, AI asset management features, and other tools to help enterprises fortify their cyber defenses and AI governance across their operations.

The broad new capabilities were unveiled recently to give business customers additional tools to better manage their expanding use of  AI agents, AI applications, and AI infrastructure as they grow their experiences with the still-evolving world of AI.

The new tools include Zscaler AI Broker, which is used to secure agentic communications through MCP (Model Context Protocol) and A2A (Agent-to-Agent) "brokers" using an integrated Agent Registry that lays out what enterprise AI agents are allowed to access, as well as new Zscaler Endpoint AI Security capabilities, which are designed to find and stop AI-related threats and risks in employee devices, browsers, plugins, extensions, and local AI tools.

Also released was Zscaler AI Access Graph, which allows enterprises to enhance AI security and agentic AI governance by better understanding and mapping AI agents, users, and identities as they communicate with AI models, applications, and data sources, according to the company.  Zscaler AI Access Graph, which was added through the company's acquisition of Symmetry Systems, enables customers to understand and enforce policies while reducing risks and unneeded access in real-time.

New features have also been added to Zscaler's AI Asset Management tools, including new capabilities to find embedded AI in SaaS and internet traffic. The tools can also now identify AI agents and MCP servers in public cloud environments while being able to find risks in agentic codebases through code scanning.

Additional controls are also now included in Zscaler's Secure Access to AI, allowing customers to expand AI interactions with prompt extractions across more than 250 GenAI apps, while adding full conversational views, support for Anthropic and OpenAI Compliance APIs, and intent-based guardrails. Enhancements to Zscaler's Secure AI Infrastructure and Apps capabilities include new AI red teaming for MCP servers, a standalone prompt hardening service, and compliance heat maps to strengthen AI governance, according to the company.

Anthony Torsiello, senior vice president of the global partner ecosystem for Zscaler, told ChannelE2E that the latest innovations are all part of Zscaler’s broader platform strategy.

"Partners can use these capabilities to help customers move from simply allowing or blocking AI tools to building a more governed AI security program," said Torsiello. "For example, partners can help customers identify shadow AI, inventory AI apps, agents, MCP servers, and AI infrastructure, map how identities and data connect to AI systems, and then enforce policy through Zscaler’s Zero Trust Exchange."

How MSPs and channel partners benefit

The new features deliver expanded sales and service opportunities for MSPs and MSSPs by helping customers with managed assessments, deployments, and ongoing operations, he said. Customers can gain help with evaluating AI risks, setting access and data policies, monitoring AI usage, detecting risky activities on endpoints, and continuously adjusting controls as AI adoption expands across businesses.

"The biggest partner value is that these capabilities give MSPs and channel partners a more complete AI security framework around which to deliver solutions and services," said Torsiello. "Instead of selling standalone point products, partners can help customers address the full AI lifecycle: discovery, access control, data protection, endpoint visibility, agent governance, red teaming, and compliance reporting."

For customers, many of whom are adopting AI faster than their internal security teams can govern it, partners can step in with consulting, deployment, managed security services, and ongoing policy management to help make it all work smoothly, he said. Channel partners can also gain new service opportunities in helping customers with AI readiness assessments, shadow AI discovery, agentic AI governance, and more, said Torsiello.

"Traditional security models were not designed for AI agents, which can act autonomously, use delegated permissions, and interact with data and systems at machine speed," he said. "As AI continues to reshape the market, our partner ecosystem will be more important than ever. The opportunity ahead is significant, and [these] announcements are another clear signal of our commitment to building with partners and investing in partner-led innovation."