Jeetu Patel, Cisco | RSAC 2026
summary
Argues that autonomous agents are digital coworkers not tools and that organizations face 10 to 1,000 agents per human scaling leading to trillions of simultaneous worker-equivalents, so security must shift from access control to action control. Describes an open-source runtime stack — DefenseClaw — built in seven days and written by AI that hooks into NVIDIA OpenShell to run skill scanners, MCP scanners and an AI bill-of-materials per agent. Calls for just-in-time, just-enough, time-limited permissions and machine-scale agentic SOC detection and interception.
AI-written summary · about
chapters (7)
- 0:00Intro
- 2:21The Scale and Impact of Autonomous Agents
- 4:56Importance of Trusted Delegation and Security
- 9:13AI Fluency and Workforce Transformation
- 12:22Simplifying AI Technologies for Broader Use
- 14:54The Role of LLMs and Agents in Technology Simplification
- 19:48Securing the Future: Innovations and Perspectives from Cisco
