What the survey found
McKinsey's 2025 State of AI survey collected responses from 1,993 participants in 105 countries between June and July 2025.
| Measure | Result |
|---|---|
| Organisations using AI in at least one business function | 88%, up from 78% |
| Have begun scaling AI across the enterprise | About one-third |
| Say AI is fully scaled | 7% |
| Report any enterprise-level profit (EBIT) impact from AI | 39% |
| "High performers", with 5% or more of EBIT attributed to AI | About 6% |
Most of those reporting a profit impact said it was less than 5% of their EBIT.
What we take from it
Using AI is no longer what sets a company apart. Almost everyone in the survey is doing it. What separates the few is getting from a pilot to something the business depends on.
In our view, that points to where to start: not with a broad AI programme, but with one process where the before and after can be measured. Questions we would ask first:
- What does this process cost today? Without a starting figure, you can't show a result later.
- Is the data ready? AI working from records that don't agree will produce answers that don't agree.
- Who checks the output? Decide early which decisions a person keeps.
A note on dates
These figures are from the 2025 edition. McKinsey's survey runs yearly, so check for a newer edition before quoting them.