Overview

Agentic AI is quickly reshaping the enterprise technology stack through autonomous, goal-driven agents that perceive, reason and act. But as adoption accelerates, organizations can’t afford to repeat the “lift-and-shift” mistakes of the cloud era, where security was too often bolted on after rapid deployment. This IDC Spotlight, presented by HCLTech in partnership with Google Cloud and Palo Alto Networks, outlines why a secure-by-design approach is essential to building resilient, scalable Agentic AI programs.

Based on IDC research and analyst insight, the paper explores the new risk landscape introduced by agentic systems, such as prompt injection, excessive agency and ecosystem/supply-chain vulnerabilities and why traditional perimeter models aren’t enough for autonomous agents that continuously interact with tools, data, users and other agents. It also highlights practical guardrails, such as bounded autonomy, strong governance, dynamic policy enforcement and runtime protections, to help teams maintain control as agent capabilities expand (including multi-agent collaboration through protocols like A2A and MCP). Grounded in IDC’s tech intelligence and market expertise, this spotlight paper provides a clear framework for aligning Agentic AI innovation with security, compliance and operational readiness, helping organizations deploy autonomous agents with confidence and measurable impact.

Key insights