The Distrust Tax: Why AI Agents Are Struggling to Deliver Value

The transition from pilot programs to full production is hitting a friction point. While 66% of companies now deploy agents in scaled environments, a growing 'Distrust Tax' emerges from the need for constant human intervention. For one-third of organizations, employees spend at least six hours every week 'bot-sitting,' an operational burden that consumes over 25% of expected productivity gains. This cycle of oversight and rework persists because companies lack the necessary proof of reliability to grant agents full autonomy.

Governance remains another significant hurdle. As 38% of organizations now manage fleets of more than 100 agents, the lack of centralized control has led to a rise in 'Shadow Agents.' Currently, only 23% of companies utilize centralized governance, with nearly half relying on informal or department-led management. This lack of visibility complicates risk management, even as 84% of buyers confirm that agent capabilities will be a defining factor in their software procurement decisions by 2027. To move beyond isolated use cases, businesses are increasingly shifting toward pay-as-you-go pricing models, signaling a departure from traditional per-seat licensing in favor of outcome-based value.