Deloitte published research on how far US organizations have got with AI agents, based on a survey of 501 leaders and 20 executive interviews. Only 15% have reached large-scale use of coordinated groups of agents, and process readiness and workforce readiness are the weakest areas reported. Most firms are adding agents on top of existing processes rather than redesigning them, largely because short payback periods are easier to justify.
What changed
Earlier discussion of enterprise AI agents focused on pilots and deployment counts rather than measured readiness of processes and workforce.
What it unlocks
Benchmarking an organization's own agent readiness against reported figures on data foundations, governance, process redesign and workforce investment at US firms.
- survey of 501 leaders involved in agentic AI strategy, plus interviews with 20 executives
- 15% have reached scaled, orchestrated multi-agent adoption
- only 16% say business processes are prepared, 5% highly prepared
- 72% cite lack of a unified data foundation as a barrier; 70% inability to trust and govern agents; 67% cost and complexity of integration
- 43% expect 'a lot' to 'extreme' job disruption within 12 to 18 months; 72% within two to three years
- 74% expect at least half of processes redesigned around agents within four years, 31% within two years
- deloitte.com2026-08-12