Key Findings

  • A report released on August 12 by global consulting firm Deloitte shows that nearly three-quarters of U.S. business leaders expect that within the next four years, about half of their business processes will revolve around AI agents.Redesign or restructure。
  • In a survey of 501 senior U.S. managers and C-level executives conducted this spring, 61% of respondents expect that most AI agents will operate with a high degree of autonomy, with humans primarily taking on supervisory roles. However, only one-fifth (20%) of leaders said their organizations are ready to transform business processes into autonomous AI agent operation.
  • Respondents pointed to multiple barriers to transformation, including "missing and poorly understood process documentation, inconsistent and fragmented data and systems, and deeply ingrained ways of working."

Deeper Insights

The report notes that AI agents are expected to boost revenue and create more meaningful work for humans, but to realize their full potential, organizations will likely need to operate in fundamentally different ways, including changes in how work is performed.

However, the survey found that 75% of business leaders believe that human-machine collaboration can create greater value than relying solely on AI-driven automation—meaning employees should guide AI agents, review outputs, verify quality, and determine when human judgment is needed.

The report explains that leaders recognize this shift will be disruptive, as new roles will emerge and job requirements will adjust to leverage human-machine collaboration. More than 40% of leaders expect a "moderate" to "significant" level of disruption over the next 12 to 18 months.

To mitigate disruption risks, Deloitte offers several recommendations, such as using layered deployment (i.e., overlaying AI agents on existing processes) as a transitional bridge rather than an end goal. The report suggests that during the layering phase, organizations can assess how human roles should evolve, who should perform the work, and how the sequence of work should be adjusted.

The report adds that organizations must also provide employees with hands-on time in AI roles, trust-building exercises to increase their confidence in working with AI agents, and clear expectations for such collaboration.

Despite issues such as technical skill gaps, employee resistance to change, and system complexity, according to a 2025 study by KPMG,leaders increasinglyview agentic AI as a tool for rethinking growth and creating new value, not just a means of cutting costs.

Additionally, a study released by the Federal Reserve Bank of St. Louis in late July found thatcorporate discussions about productivityare increasingly framed through the lens of AI, with U.S. executives generally optimistic about the technology's future impact.

Given these expectations, a July report from Culture Amp noted that to maintain organizational trust, HR professionalsmust move beyond task-level AI applicationsand shift toward agentic workflows. However, a global survey found that only about one-third of HR professionals say they have made this shift; the platform found that most organizations still use AI as intelligent assistants rather than autonomous agents operating within defined permissions.