Agentic AI adoption has stalled at the experimentation stage for the vast majority of US companies, with only 15% having achieved scaled, orchestrated multi-agent deployments, according to a survey of 501 senior business leaders published by Deloitte in August 2026. Even within that leading group, the gaps are stark.
The survey, fielded between April and June 2026 across five industries including financial services, life sciences, and technology, media and telecommunications, found that 42% of organisations are still testing small numbers of agents and 43% are expanding deployments across functions. Only the remaining 15% have moved to full, orchestrated production.
Only 15% Have Scaled, and Even They Face Gaps
Reaching scale has not solved the underlying readiness problem. According to the Deloitte Insights report on agentic transformation, fewer than half (46%) of those scaled adopters believe their business processes are prepared for agentic adoption. Across all organisations surveyed, only 5% say their business processes are highly prepared.
The readiness figures broken down by function make the challenge concrete. Only 36% of leaders say their vision and strategy is ready for agentic AI, 34% point to technology infrastructure, 32% to their data foundation, and just 26% to risk, security, and governance. Only 1 in 5 businesses say their workforce is ready.
Many scaled adopters are also applying their deployments in low-risk, low-return-on-investment applications, according to Deloitte, suggesting that reaching scale on the scoreboard does not necessarily mean capturing meaningful business value.
Barriers to Agentic AI Adoption Run Deeper Than Technology
The survey identifies three leading barriers to scaling: a lack of a unified, accessible data foundation (cited by 72% of leaders), an inability to trust and govern agents (70%), and the cost and complexity of integration (67%).
Those governance concerns are reinforced by a separate, larger Deloitte study. The 2026 State of AI in the Enterprise report, which surveyed 3,235 IT and business leaders across 24 countries, found that only 21% of respondents say their organisations have a mature governance model in place for agentic AI, with approximately 80% currently lacking mature governance capabilities.
The same multicountry study found that by 2027, 74% of respondents expect their companies to be using AI agents at least moderately, with 23% expecting extensive use and 5% anticipating full integration as a core component of business operations.
Beyond governance, the survey points to a structural problem: companies are layering AI onto legacy processes rather than rebuilding those processes for an agentic world. Deloitte’s research suggests layering can serve as a bridge, but it cautions that process redesign needs to become a core organisational capability. Only 31% of businesses expect to have redesigned and rebuilt their processes around AI agents by 2028, yet 74% expect that transformation by 2030.
The Workforce Gap Is Slowing Progress
Half of the leaders surveyed said their organisations are not investing enough in workforce transformation to support successful agentic AI adoption. Nearly two-thirds are reevaluating their business models in light of advances in agentic AI, yet only half have a clear view of what their future operating model will look like.
The disruption to jobs is expected to be substantial. Some 43% of business leaders anticipate major job disruption from agentic deployments as routine and structured tasks become autonomous. In response, 71% of organisations are working on baseline AI literacy programmes and 65% are actively upskilling and reskilling employees in roles likely to be affected.
By 2030, 61% of leaders believe most of their processes will be powered by AI agents operating with little to no human involvement. Currently, only 16% say their processes are prepared for that level of autonomy.
A Large and Growing Market, With Real Cancellation Risk
The market opportunity is substantial. Deloitte’s TMT Predictions 2026 report estimates the autonomous AI agent market could reach $8.5 billion by 2026 and $35 billion by 2030. If enterprises improve agent orchestration, the projection could rise by 15% to 30%, reaching as high as $45 billion by 2030.
Execution risk, however, is real. Deloitte’s TMT Predictions report cites a Gartner estimate that more than 40% of today’s agentic AI projects could be cancelled by 2027 due to unanticipated cost and complexity.
That risk sits alongside a broader enterprise shift. A separate Deloitte report found that AI tools are now available to the workforce of about 60% of surveyed organisations, according to the announcement of Deloitte’s Google Cloud agentic transformation practice, signalling a move from pilots to enterprise scale. Meanwhile, Gartner, cited within a Deloitte Tech Trends 2026 analysis, predicts that by 2028, 33% of enterprise software applications will include agentic AI, up from less than 1% in 2024, and that at least 15% of day-to-day work decisions will be made autonomously by agents.
The 2028 redesign deadline, when only 31% of businesses expect to be ready, is now less than two years away.