Enterprise Report · The AI Index
Thank you for reading this post, don't forget to subscribe!
Home / Reports / Deloitte State of AI 2026
Agents Are Scaling Faster Than the Guardrails
Deloitte’s State of AI in the Enterprise 2026 finds AI moving from pilots to scaling — 74% of leaders expect to use AI agents by 2027, yet only 21% have a mature governance model for agentic AI. Here is what the survey of 3,235 leaders says, and why the gap matters.
By The AI Index · Updated Jun 23, 2026 · 5 min read
Key takeaways
- By 2027, 74% of respondents expect to use AI agents at least “moderately,” 23% “extensively,” and 5% fully integrated as a core component of operations. (Deloitte)
- 85% expect to customize agents to fit their business’s unique needs. (Deloitte)
- Only 21% say their organization has a mature governance model for agentic AI — roughly 80% lack clear decision boundaries, real-time monitoring, and audit trails. (Deloitte)
- Deloitte’s framing: AI is shifting from experimentation to scaling, but agents are “scaling faster than their guardrails.” Companies that succeed take a measured approach. (Deloitte)
From ambition to activation
Deloitte’s 2026 edition of its State of AI in the Enterprise report carries the subtitle “From Ambition to Activation: Organizations Stand at the Untapped Edge of AI’s Potential.” It is based on a survey of 3,235 business and IT leaders across 24 countries and six industries, conducted in August and September 2025. The headline shift: AI is moving out of pilots and experimentation and into scaling — but the move is uneven.
Agents go mainstream by 2027
The clearest signal is agent adoption. By 2027, 74% of respondents expect their companies to use AI agents at least “moderately,” 23% expect to use them “extensively,” and 5% expect to fully integrate agents as a core component of operations, according to Deloitte’s survey. And this is not off-the-shelf adoption: 85% expect to customize agents to fit their business’s unique needs, signalling that agents are being treated as core operational infrastructure rather than bolt-on tools.
The governance gap
The problem is what comes with that scale. Only 21% of respondents say their organization has a mature governance model for agentic AI — meaning roughly 80% lack the basics: clear decision boundaries, real-time monitoring, and audit trails. As Deloitte puts it, agents are “scaling faster than their guardrails.” Autonomous systems that take actions — not just generate text — raise the stakes when oversight is thin.
— Deloitte, State of AI in the Enterprise 2026
What it means
Deloitte’s prescription is a measured approach. The companies succeeding with agents start with lower-risk use cases, build governance alongside deployment, and scale deliberately rather than racing to integrate autonomy everywhere at once. The gap between the 74% planning to use agents and the 21% with mature governance is, in Deloitte’s reading, the central enterprise-AI risk of the next two years — and the most addressable.
The numbers in full
| Metric | Figure |
|---|---|
| Expect to use AI agents at least “moderately” by 2027 | 74% |
| Expect to use AI agents “extensively” by 2027 | 23% |
| Expect to fully integrate agents as a core component of operations | 5% |
| Expect to customize agents to their business’s needs | 85% |
| Have a mature governance model for agentic AI | 21% |
| Business and IT leaders surveyed | 3,235 |
| Countries / industries covered | 24 / 6 |
Frequently asked
How many companies plan to use AI agents by 2027?
In Deloitte’s State of AI in the Enterprise 2026 survey, 74% of respondents expect their companies to use AI agents at least “moderately” by 2027, 23% expect “extensive” use, and 5% expect to fully integrate agents as a core component of operations.
How many organizations have mature agentic-AI governance?
Only 21% of respondents say their organization has a mature governance model for agentic AI — meaning roughly 80% lack clear decision boundaries, real-time monitoring, and audit trails, according to Deloitte.
What does Deloitte recommend for scaling AI agents safely?
Deloitte advises a measured approach: start with lower-risk use cases, build governance alongside deployment, and scale deliberately — because agents are “scaling faster than their guardrails.”
Primary source: Deloitte
• Deloitte — AI agents scaling faster than guardrails
• PRNewswire — survey announcement