Cisco & AppOmni Secure SaaS AI Agents: New Protections

Cisco AI Defense and AppOmni are integrating to provide new security for AI agents in SaaS platforms. This means organizations can now protect AI agents running in services like Microsoft 365 Copilot and ServiceNow Now Assist. Here's how it works: AppOmni AgentGuard monitors AI agent interactions within SaaS environments. It then sends these interactions to the AppOmni cloud analysis engine, which calls Cisco AI Defense for guardrail evaluation. AI Defense applies configured policies and determines whether to allow, block, or terminate the interaction. This system offers real-time protection against various attacks, including prompt injection, data exfiltration, and harmful content. These protections are enforced before the SaaS agent can act and are updated by Cisco’s AI security research team. Security teams can also define custom guardrails to address specific threats relevant to their agents, industries, and use cases. This integration provides a uniform security policy across all AI agents, whether internally built or running in SaaS. This aims to ensure SaaS AI agents are no longer a security exception.