Sixty percent of companies worldwide have adopted AI tools — but only 25% have pushed more than 40% of their pilots into production. That's the headline finding from Deloitte's survey of 3,235 C-suite leaders across 24 countries: enterprise AI is long on ambition and short on activation.
At a Glance
Survey scope: 3,235 executive leaders across 24 countries and 6 industries (conducted August–September 2025)
Key finding: Only 25% of companies have moved 40%+ of AI pilots into production — but 54% say they could get there within 6 months
Signal of change: Employee AI access expanded 50% in one year (40% → 60%); 74% of companies plan to deploy agentic AI within 2 years
Achilles' heel: Only 34% have redesigned their business model around AI; 84% haven't touched job design at all
What Is It?
This is the 'The State of AI in the Enterprise: The Untapped Edge' report, published by the Deloitte AI Institute in January 2026 at Davos. It's the seventh edition of this annual series, and this year's subtitle — "Ambition to Activation" — tells you everything you need to know about the current moment.
The core message is simple: most companies have an AI strategy, but very few have reached the "activation" stage where that strategy actually generates business value.
The AI execution gap by the numbers
- 25% — companies that have moved 40%+ of AI pilots into production
- 34% — companies fundamentally redesigning their business model with AI
- 37% — companies using AI only at the surface level, barely changing existing processes
- 84% — companies that haven't redesigned jobs to match AI capabilities
- 21% — companies with a mature governance model ready for agentic AI
Around the same time, McKinsey released its 2026 AI Trust Maturity Survey. Across roughly 500 companies, only about 30% had reached maturity level 3 or above in strategy, governance, and agentic AI controls. That's a near-perfect match with what Deloitte found.
Deloitte Korea published the same report in Korean with commentary on local implications. Jae-min Bae, leader of Deloitte Korea's AI Integration Services, put it plainly: "A company's competitive edge depends on how deeply AI is embedded into the organization."
What Changes?
Until recently, enterprise AI conversations centered on whether to adopt AI at all. Starting this year, the question has shifted to: we've adopted it — so why aren't we seeing results? Here's how the Deloitte report frames the shift.
| Area | Before (2024–2025) | Now (2026–) |
|---|---|---|
| Focus | Running AI pilots | Production deployment · business model redesign |
| AI access | Fewer than 40% of employees | Expanded to 60% of employees (up 50% in one year) |
| Success metrics | Productivity gains | Revenue growth · strategic differentiation (only 20% have tied AI to revenue) |
| AI type | Mostly generative AI | Agentic AI + physical AI + sovereign AI spreading simultaneously |
| Talent strategy | Hiring AI specialists | AI fluency training for all employees (53% say it's the top priority) |
| Governance | Regulatory compliance focus | Reframed as an enabler of business growth |
| Vendor strategy | Global big-tech dominated | 77% now factor country of origin into vendor selection |
The most significant shift is the rise of agentic AI. Seventy-four percent of companies plan to deploy autonomous agents within two years — but only 21% have governance models in place to manage them. McKinsey found the same thing: security and risk concerns are the biggest barrier to scaling agentic AI.
Physical AI is also moving fast. Fifty-eight percent of companies are already using it at some meaningful level, with 80% adoption projected within two years. Manufacturing, logistics, and defense are leading the charge, and the Asia-Pacific region is out front on early adoption.
Getting Started
Putting the Deloitte and McKinsey findings together, here's a practical roadmap for closing the AI ambition-to-execution gap.
- Break pilot fatigue with clear graduation criteria
Set a go/no-go threshold for every AI pilot: within 6 months, it either moves to production or gets shut down. In Deloitte's survey, 54% of companies said they could hit the 40% production threshold within 6 months — the bottleneck isn't capability, it's the lack of a forcing function. Without clear criteria, you just keep running experiments. - Design KPIs that go beyond productivity
Sixty-six percent of companies report productivity gains from AI, but only 20% have connected that to actual revenue growth. Don't just track efficiency metrics like time saved or cost reduced — build in strategic KPIs: revenue attribution, new product launch speed, customer experience scores. That's what turns AI from a cost play into a growth driver. - Start redesigning jobs, not just training people
Eighty-four percent of companies still haven't redesigned jobs around AI. AI fluency training is necessary but not sufficient. You need to rethink role structures, workflows, and career paths. Define new job titles concretely — AI operations managers, human-AI interaction specialists — rather than leaving it vague. - Build agentic AI governance before you deploy agents
Get the governance framework in place before the agents go live. Start with low-risk use cases, build your governance muscle, then expand from there — that sequencing consistently produces better outcomes. McKinsey data backs this up: organizations with clear accountability structures score 2.6 on maturity versus 1.8 for those without. - Audit your sovereign AI position
Seventy-seven percent of companies now factor country of origin into vendor selection, and 58% are prioritizing local vendors in their AI stack. Take a hard look at your AI infrastructure from three angles: data sovereignty, regulatory compliance, and supply chain risk. Make sure you're not strategically dependent on vendors you can't control.
Deep Dive Resources
Deloitte 2026 Report (Full Text)
Based on surveys of 3,235 executives across 24 countries, the full report covers seven key findings in depth — including agentic AI, physical AI, and sovereign AI. Free PDF download available.
McKinsey 2026 AI Trust Maturity Survey
Rates 500 companies on responsible AI maturity across five dimensions. The finding that strategy and governance gaps outpace technical capability gaps maps almost exactly onto what Deloitte found — useful cross-validation.
Deloitte Korea Report (Korean)
The global survey findings with added commentary on implications for the Korean market. Focuses specifically on what the rise of agentic AI and sovereign AI strategy means for domestic companies.
