AI agents are racing into the enterprise, but trust isn’t keeping pace.
That’s the headline from new research conducted by Harvard Business Review, in partnership with Workato and AWS. And the findings confirm what many CIOs and CTOs feel every day: organizations are excited about agentic AI, but few believe their foundations are mature enough to let AI agents operate at the core of their business.
A recent Fortune analysis of the study underscores the urgency: only 6% of companies fully trust AI agents to autonomously run core business processes, yet 86% plan to increase their investment in agentic AI over the next two years
This widening gap between ambition and readiness is the defining challenge of enterprise AI adoption.
And it’s exactly where enterprise orchestration and the Enterprise MCP come in.
AI Agents Are Coming Fast—But Core Systems Aren’t Ready
The study shows explosive interest in agentic AI, but deep hesitation around letting these agents act without supervision:
- 43% trust AI only for limited or routine tasks
- 39% restrict AI agents to supervised or noncore processes
- Only 20% believe their infrastructure is ready, and just 15% feel their data and systems are prepared for agentic AI at scale
A readiness index in the report shows that:
- 27% of organizations qualify as leaders,
- 50% as followers,
- 24% as laggards—not yet structurally prepared for agentic AI integration
Leaders aren’t succeeding by experimenting with more agents—they’re succeeding because they’re investing in the foundation that makes agents viable, safe, and predictable.
The Trust Gap Is an Orchestration Problem
Executives were clear about what’s holding AI agents back:
- 31% cite cybersecurity and privacy concerns
- 23% cite data output quality
- 22% cite unready business processes
- 22% cite limitations in technology infrastructure
These aren’t “AI” problems. They’re enterprise architecture problems.
The study points to a solution many organizations are already pursuing:
Enterprise Orchestration.
More than 74% of organizations are working on or planning to implement enterprise orchestration to prepare for agentic AI—because orchestration provides the connective tissue agents need to safely operate across systems, data, and workflows.
This is why the foundation of Enterprise MCP is so important. It’s a governed, secure, AI-ready orchestration layer that turns systems, data, and processes into consumable services agents can safely use.
Download the Full Harvard Business Review Report
The full research report—“Enterprise Agentic AI: The Foundations of Trust”—is now available.
It includes insights from over 600 senior IT and business leaders on:
- The current state of AI agent adoption
- What separates leaders from laggards
- The key risks enterprises are navigating
- The rise of enterprise orchestration as the new architectural standard
- How organizations are preparing for autonomous AI in core workflows