Innovation Brief · IB-2026-01a · January 2026
Trust as Infrastructure
Building secure systems in an agentic era
Agents act autonomously across systems and can modify their own behaviour. Trust stops being a perception problem and becomes infrastructure.
- Authors
Dave WrightChief Innovation Officer
- Length
- 16 pages
- Themes
- Trust & Security, The Future of AI, AI Strategy & Maturity
Summary
- Agentic AI is forcing enterprises to reckon with their existing trust and security infrastructure.
- Organizations can embed trust into their AI architecture through visibility and control, robust monitoring and identity infrastructure, active human oversight and cross-functional governance councils, and advanced AI security capabilities.
- Unlike human users, AI agents operate autonomously across systems and can modify their own behavior, requiring fundamentally different security approaches.
- A solid trust architecture lets organizations deploy AI capabilities that competitors cannot risk. Those who wait will scramble to retrofit security after breaches occur.
From the brief
In enterprise IT, trust has traditionally meant confidence that systems will behave reliably, securely and as intended. Agentic AI is evolving that definition.
In the agentic era, trust requires both reliability — ensuring agents execute within defined boundaries — and intent: continuously verifying they act with transparency into their decisions and oversight of their impact.
Read the full report