Key Takeaways

  • KPMG's latest AI quarterly pulse survey shows that the deployment scale of coordinated AI agents is expanding, a shift that brings newcost visibility challenges。
  • The proportion of organizations orchestrating multiple AI agents across workflows has doubled from 9% in the previous survey to 18%. Respondents indicated that agents are used to align common goals and success metrics across functions (64%), support joint decision-making (49%), and automate cross-functional workflows (48%).
  • "AI agents are simultaneously transforming operating models and economic models," said Rahsaan Shears, KPMG's lead for enterprise AI transformation, in a press release. "As organizations move from isolated deployments to coordinated enterprise-wide applications, good governance is the key to connecting scale, performance, and value."

Deep Insights

The survey results reflect that enterprise AI adoption is moving from the experimental stage to a more mature phase—agents "help unify execution and improve consistency," the report states.

Although the proportion of organizations deploying AI agents remained roughly flat compared to the previous survey (53% versus 55%), enterprises are managing increasingly complex deployments that span teams, systems, and decision points, KPMG found.

This shift is exposing shortcomings in how enterprises manage AI spending.

While many organizations have implemented governance tools (such as dashboards and approval processes), most still lack real-time, end-to-end visibility into AI-related costs, KPMG found.

Only 26% of organizations have real-time visibility into the costs of running AI at scale, while 35% of leaders say that AI cost management and financial literacy—including understanding usage-based pricing models such as token and inference costs—remain major obstacles.

KPMG notes that this readiness gap is becoming increasingly urgent as organizations move from isolated AI use cases to coordinated enterprise-wide agent deployments.

"Organizations are facing an emerging challenge: understanding and managing the costs of operating AI at scale," the report states.

The survey was conducted from April 28 to May 25, with a sample covering 204 U.S. enterprise leaders from organizations with annual revenue of at least $1 billion.