newsfilter
← Latest News
PR Newswire·

Cisco Unveils Frontier AI Red Teaming and Agent Observability to Secure the Agentic Enterprise

CSCO
  • Cisco AI Defense is integrating Anthropic's Claude Mythos to provide advanced adversarial reasoning for adaptive red teaming of AI agents.
  • New Splunk Tokenomics features will connect AI agent costs directly with return on investment (ROI) in real time to optimize spending.
  • Cisco IQ now correlates live threat detection data with exposed network infrastructure at machine speed to accelerate remediation.
  • The combined innovations create a continuous trust loop for securing, governing, and observing complex enterprise AI deployments.
  • Claude Mythos in Cisco AI Defense generates multi-step, contextual attack paths while operating exclusively on registered, authorized targets with named operator attribution.
  • Splunk Agent Observability introduces Eval-G1 evaluation models that execute evaluations at 98% lower cost than commercial LLMs with 2.5x faster response times.
  • The Evaluation Tuner capability allows teams to create fine-tuned evaluator models from agentic traces to enforce quality standards.
  • Cisco Data Fabric integration enables ingestion of Cisco IQ vulnerability data into Splunk, allowing AI agents to correlate threats with network exposure.
  • Model Context Protocol (MCP) & ServiceNow connectors automatically translate exposure findings into asset-linked problem records to coordinate remediation.
  • Adaptive red teaming with Claude Mythos is coming soon for select enterprises, while Tokenomics enters private preview in December 2026.
  • Evaluation models & the evaluation tuner are generally available starting in November 2026.
  • Tokenomics will reach general availability in the AWS Marketplace in early 2027.
Company
Cisco
Keywords
Cisco