The AI Value Gap: 91% of Professionals Say Firms Fail to Realize AI Potential

The vast majority of professional organizations are failing to capture the actual value of artificial intelligence. According to a new report from Thomson Reuters, 91% of employees believe their organizations fall short of reaching the value AI can deliver.

Based on a global survey of 1,800 professionals across multiple sectors, the findings highlight a disconnect between corporate ambition and operational reality. While firms are deploying AI, 41% of professionals using the technology at work report a lack of access to high-quality tools. Furthermore, the existence of a formal plan does not guarantee execution; 35% of professionals in firms with a named AI strategy say that strategy is not visible in their daily work.

The Era of Tool Blast

This discrepancy stems from a period of eager experimentation that began roughly 18 months ago. Many firms have entered a phase of "tool blast," deploying a fragmented array of frontier, private, and domain models without redesigning the underlying business processes. This has created a hyper-fragmented landscape where employees are often confused about which tools to use and how to integrate them into existing workflows.

Steve Lucas, CEO of Boomi, described the current state of play in AI as "hyper-fragmented and hyper-fractured." This lack of cohesion makes it difficult for leadership to track productivity gains. Kirsty Roth, COO at Thomson Reuters, noted that when firms are unclear on which tools staff should use, it becomes very hard to see the uplift, while software costs continue to rise significantly.

The Requirements for Production

To bridge this gap, professionals are signaling a need for more rigorous, production-level standards over general experimentation. The survey found that users prioritize three critical requirements: safeguarding confidential data (96%), grounding outputs in authoritative content (94%), and producing explainable reasoning (90%).

Without these safeguards, the risk of failure increases. Gartner predicts that 40% of enterprises will demote or decommission autonomous AI agents by 2027, citing governance gaps that are often only identified after production incidents occur.

Why the Gap Persists

The "value gap" indicates that simply purchasing AI licenses is insufficient for productivity gains. The industry is currently clashing with rising IT costs and "token maxing" in the absence of tangible ROI. Success now requires a fundamental shift in change management—moving away from providing a "playground" of tools toward the redesign of business processes.

Firms that fail to transition to specific, well-grounded use cases risk wasting significant capital on software that provides no measurable uplift. The path forward requires shifting from unfocused experimentation to production-level explorations that solve specific business problems.

What's Next

As the initial hype cycle fades, the focus is shifting toward governance and integration. The industry must now determine how to move AI from a peripheral assistant to a core component of professional workflows. Whether firms can successfully implement the data safeguarding and explainability requirements demanded by their staff will likely determine who survives the transition from experimentation to actual value.