85% of enterprises are running AI agents. Only 5% trust them enough to ship.

2026-04-24 · Source: VentureBeat · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cybersecurity & Data Privacy · Depth: Advanced, medium

Summary

Cisco President and Chief Product Officer Jeetu Patel revealed at RSA Conference 2026 that 85% of enterprises are piloting AI agents, but only 5% have moved them into production, attributing this 80-point gap to a lack of trust architecture. He emphasized that "trusted delegation" is crucial for market dominance. Cisco's response includes new products like AI Defense Explorer Edition, Agent Runtime SDK, and the LLM Security Leaderboard. Notably, Cisco integrated its Defense Claw open-source framework with Nvidia's OpenShell within 48 hours, enabling automatic security enforcement for agents. Patel also disclosed a mandate for Cisco's 90,000-person engineering organization to achieve 70% AI-built products by the end of 2027, with AI Defense already 100% AI-coded.

Key takeaway

For CTOs and VPs of Engineering grappling with AI agent deployment, your focus must shift from mere delegation to "trusted delegation." Implement robust security architectures that integrate policy enforcement at build-time and establish clear agent behavioral baselines. Without distinguishing agent-initiated actions from human ones in your SIEM, your identity layer security remains incomplete, exposing your organization to significant, irreversible risks.

Key insights

Establishing a robust trust architecture is critical for scaling AI agent adoption from pilot to production in enterprises.

Principles

• Trusted delegation prevents catastrophic agent actions.
• Speed in security integration is paramount for agent frameworks.
• AI dexterity drives significant productivity gains.

Method

Cisco's approach involves protecting agents from external threats, protecting systems from agents, and detecting/responding at machine speed, integrating security at build-time via SDKs and open-source frameworks like Defense Claw.

In practice

• Use AI Defense Explorer Edition for red-teaming agent workflows.
• Integrate security services at container launch via OpenShell.
• Map agent delegation chains to identify human approval gaps.

Topics

• AI Agents
• Trust Architecture
• Enterprise AI Adoption
• Cisco Security Products
• Defense Claw

Best for: CTO, VP of Engineering/Data, Executive, AI Security Engineer, Director of AI/ML, MLOps Engineer

Related on AIssential

• OpenClaw and Claude Opus 4.6: Where is AI agent security headed? — AI agents and rapid development necessitate stringent security, zero-trust principles, and comprehensive inventory management.
• Agentic Trust: Securing AI Interactions with Tokens & Delegation — Securing agentic AI systems requires robust identity, authentication, authorization, and secure propagation mechanisms.
• Palo Alto Networks' Bid to Secure the Agentic Enterprise — Centralized AI gateways are crucial for securing and governing autonomous AI agents in enterprise environments.
• Secure agentic AI end-to-end — Securing agentic AI requires ambient, autonomous security woven into every layer of the AI estate.
• Identity-first AI governance: Securing the agentic workforce — Autonomous AI agents require distinct, governed identities within enterprise IDPs to mitigate significant security and compliance risks.
• Zero Trust for AI Agents — Securing autonomous AI agents requires adapting Zero Trust principles to dynamic, multi-layered threats.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by VentureBeat.