India, Nov. 10 -- As enterprises move AI agents into core business operations, securing them is becoming more complex. These agents can connect to tools, access sensitive data and carry out multi-step tasks, creating new risks that traditional testing may not fully uncover. At Cisco Live APJC, Cisco announced updates across AI Defence, Splunk Agent Observability and Cisco IQ to help enterprises secure, monitor and govern these systems.

The key addition is Claude Mythos from Anthropic, which will support advanced AI red teaming within Cisco AI Defence. The move reflects a wider shift in enterprise AI security: organisations need to test not only whether an AI model works, but also how it behaves when faced with carefully designed attacks....