---
id: research-mckinsey-state-of-ai-2025
category: research
date_observed: 2025-11
date_added: 2026-04-24
source_quality: primary
status: active
affected_levels: [1, 2, 3, 4]
affected_functions: [all]
tags: [adoption-stats, value-capture-gap, high-performers, workflow-redesign]
sources:
  - url: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
    title: "The State of AI in 2025: Agents, Innovation, and Transformation (McKinsey, November 2025)"
  - url: ../../Resources/the-state-of-ai-in-2025-vf.pdf
    title: McKinsey State of AI 2025 (local PDF)
maintainer: rafa
---

# McKinsey — The State of AI in 2025

## Summary
McKinsey's annual AI survey (1,993 respondents across 105 nations, fielded June 25–July 29, 2025). Reports 88% organizational adoption of AI in at least one function (up from 78% a year prior), but only 39% reporting any EBIT impact and roughly 6% qualifying as "high performers" (>5% EBIT from AI + significant value). The report's central theme is the widening gap between adoption and value capture, with workflow redesign, human-in-the-loop validation, and senior-leadership ownership differentiating high performers.

**Correction note:** An earlier annotation cited "~33% (2023) → ~65% (2024)" adoption figures as McKinsey. That figure was a misattribution from a historical time series and should not be used. The verified 2025 adoption figure is 88%.

## Why it matters for the CMM
Empirical spine for Section 5 (where organizations are now). The adoption/value-capture gap — 88% adoption vs. 6% high performers — is the strongest single-source articulation of the CMM's core claim: maturity is not tool adoption. The high-performer differentiators (workflow redesign, human-in-the-loop, senior leadership ownership) map cleanly onto Levels 3 and 4 of this model. The agentic-AI figure (22% scaling anywhere, no more than 10% in any single function) also anchors the Level 3+ prevalence estimate.

## Claims citable from this observation
- 88% of organizations use AI in at least one function (2025); 78% a year prior
- Only 39% report any EBIT impact from AI
- ~6% qualify as high performers (>5% EBIT impact plus "significant value")
- 66% have not yet begun scaling AI across the enterprise
- "Meaningful enterprise-wide bottom-line impact from the use of AI continues to be rare." (verbatim, McKinsey)
- Human-in-the-loop validation: 65% of high performers have it vs. 23% of others
- "High performers are nearly three times as likely as others are to say their organizations have fundamentally redesigned individual workflows in their deployment of AI." (verbatim, McKinsey)
- Senior leadership ownership: high performers 3.0x more likely (48% vs. 16%)
- Only 22% scaling agents anywhere; no more than 10% in any single function
- Survey fielded with 1,993 respondents across 105 nations, June 25–July 29, 2025

## Related observations
- [menlo-state-of-genai](menlo-state-of-genai.md) — complementary adoption baseline (16% true agents)
- [accenture-art-of-ai-maturity](accenture-art-of-ai-maturity.md) — complementary Achievers/Experimenters frame
- [openai-state-of-enterprise-ai](openai-state-of-enterprise-ai.md) — frontier-vs-median gap is the vendor-level view of the same pattern