Enterprise AI Stalls When Departments Won't Work Together
Functional silos are blocking AI value at 82% of companies, according to new research from IBM Institute for Business Value. The survey of 2,000 senior technology and business leaders across 16 countries found that organizations are already shifting toward a different model: agentic AI operating systems built around workflows instead of departments.
This matters directly to operations leaders. The traditional structure-where departments operate independently-was never designed for AI agents that need to move work across functions to complete end-to-end tasks. Breaking those boundaries is becoming necessary infrastructure, not optional.
The Operating Model Is Changing
IBM calls the emerging alternative the "interconnected enterprise." In this model, AI agents handle routine execution and operational decisions. Human experts focus on judgment, oversight, ethics, exceptions, and trade-offs.
The shift is already underway. Seventy-five percent of executives say AI will significantly redefine global service delivery by the end of 2026. Sixty-one percent are actively dismantling boundaries to build a more unified digital culture.
Fifty-five percent of organizations are already developing or deploying an agentic AI operating model.
Workflows Become the Unit of Performance
This changes how work gets measured and priced. Fifty-six percent of executives expect next-generation delivery models to be priced around measurable outcomes-customer impact and risk exposure-instead of transactions or headcount.
That reframes workflows as business-value engines where AI agents execute across systems and partners. Human experts focus on oversight and decisions that directly affect enterprise outcomes.
Data Must Connect Across the Enterprise
Autonomous AI depends on data that is accessible, consistent, and connected. Seventy-seven percent of leaders are investing in data quality and orchestration. Fifty-nine percent now see interoperable AI capabilities across business functions as a priority.
These investments protect agentic systems from making decisions based on fragmented data. Complete and reliable information becomes a safeguard, not a luxury.
Digital Twins Control Autonomous Workflows
Enterprise digital twins-real-time models that mirror operations-are emerging as the control layer for agentic systems. Eighty-two percent of executives say a unified digital twin control plane is essential for autonomous operations.
Seventy-two percent expect real-time simulation and scenario analysis to become core features of next-generation service delivery.
Six Capabilities Determine Scale
IBM identifies six capabilities that matter: change management readiness, AI governance, data governance, real-time shared data integration, system interoperability, and financial integration.
When all six mature together, organizations are 5.4 times more likely to adopt autonomous workflows. Change management and AI governance rank as especially critical. The shift is as much organizational as technological.
For operations professionals, this means the work of breaking silos is foundational. Without it, agentic AI projects remain isolated automation efforts instead of becoming enterprise-wide value engines. Learn more about AI for Operations and AI Agents & Automation to understand how these systems work in practice.
Your membership also unlocks:
