Enterprise AI in 2026: the long road from pilot to production

Executive summary

Use is close to universal in executive surveys, but scale is not. Nearly nine in ten respondents to McKinsey’s 2026 survey say their organisation uses AI regularly in at least one function, yet only 44% say AI is scaling across the enterprise.[1]
Financial impact has stopped rising. 37% of respondents attribute any EBIT impact to AI, essentially unchanged from 2025, and the high performers who attribute 5% or more of EBIT to AI remain about 6% of the sample.[1] BCG puts the share of companies achieving value at scale at 5%.[8]
Agents are reaching production, mostly in large organisations. Agent scaling at large organisations rose from 27% to 40% in a year, while smaller organisations stayed at 22%.[1] Gartner still expects more than 40% of agentic AI projects to be cancelled by the end of 2027.[10]
Spending is growing faster than returns. Gartner forecasts worldwide AI spending of $2.7 trillion in 2026, up 49.5%, and places generative AI “firmly in the Trough of Disillusionment”.[5]
Data foundations are the binding constraint. Gartner’s 2026 CDAO survey finds 85% of data and analytics practices are not architected for scale.[13] It also predicts that 60% of AI projects unsupported by AI-ready data will be abandoned through 2026.[12]
The EU AI Act timetable has moved. The AI Omnibus entered into force on 27 July 2026. Stand-alone high-risk obligations now apply from 2 December 2027 and product-embedded ones from 2 August 2028, while transparency rules have applied since 2 August 2026.[15][16]

Key findings

0144% of organisations say AI is scaling across the enterprise, up from 38% a year earlier

0237% attribute any EBIT impact to AI, unchanged from 2025, and high performers remain about 6%

03Worldwide AI spending is forecast at $2.7 trillion in 2026, up 49.5%

Organisations scaling AI agents20252026

Large organisations

27%

40%

Smaller organisations

22%

Source: McKinsey, 2026

Organisations where AI is scaling across the enterprise

2025

38%

2026

44%

Source: McKinsey, 2026

What leaders should do

Fund products and outcomes rather than annual projects.
Set baselines before launch and track value after it.
Stop low-value initiatives early through stage gates.

In the full report

Adoption: wide, shallow and uneven
Agents: early production, heavy attrition
Where value shows up, and where it does not
Spend: rising fast and shifting towards embedded and bought AI
Data readiness and governance are the bottleneck
Skills and the operating model
EU AI Act: the timetable as it stands in October 2026
What to do in 2027
Sources

Methodology

ConfX research brings together our own analysis, what we hear from senior leaders at ConfX summits, executive roundtables and briefings, and the latest published research from regulators, academic institutions, industry bodies and research firms. Analysis completed on 9 October 2026.

Every figure is attributed to its source. Samples, regions and definitions differ between studies, so figures from different sources describe different groups and should not be combined.