Gartner IT Symposium/Xpo™ Keynote

Many organizations equate growth in AI activity with growth in business value. However, Gartner finds the average organization increased its number of AI use cases by more than 400% in a single year, yet the majority of their AI projects are failing to deliver positive ROI. Activity without value is simply bad momentum.

Instead of launching new projects, organizations should focus more on creating repeatable advantages. A successful initiative should create reusable capabilities, trusted governance, expertise and data assets that improve future results.

The goal is not to create more AI activity but to create outcomes where one success makes future successes easier, faster and more valuable. Organizations that repeatedly reinvent the wheel create more effort and more complexity, but not necessarily more value.

The Gartner Value-Risk Flywheel is shaped by two forces: momentum and drag. Momentum accelerates outcomes, while drag slows them down. However, neither force is inherently good or bad. Good momentum compounds value through trust, reuse, work redesign and scaling proven successes. Bad momentum creates activity, duplication and complexity without creating advantage.

The same principle applies to drag. Good drag protects value through governance, cybersecurity, accountability and spending discipline. Bad drag creates bureaucracy, bottlenecks and rework without reducing risk.

As AI becomes embedded across people, agents, models and platforms, accountability becomes harder to see and harder to manage. To address this challenge, Gartner introduced the concept of an AI Central Bank. Rather than governing AI through disconnected functions, the AI Central Bank acts as a common control plane that brings together accountability and governance, technology integration, cost economics and shared intelligence. This approach helps organizations improve visibility, coordinate decisions and balance value with risk across the enterprise. It also helps address the invisible chains of accountability emerging as AI systems become more autonomous.

Yet technology and governance alone are not enough, people remain essential. People are the first responders to AI failures, the last line of defense against dark AI and the entrepreneurs who discover new ways to create value. CIOs should make “people to the rescue” their mantra, positioning people as both value drivers and risk mitigators in an increasingly AI-shaped enterprise.