Марія Бровінська AI Eng 12 November 2025, 09:01

How companies will actually use AI in 2025 - six insights from a McKinsey report

The top 6 most important insights from the McKinsey report were collected by Generative AI Consultant Alexey Minakov.

High interest in artificial intelligence agents

62% of respondents say their organizations are at least experimenting with AI agents. But the use of agents is not yet widespread: most companies still have agents in 1-2 functions, and only up to 10% are scaling them systemically.

Most organizations are still in the experimental or piloting stage

88% of organizations already regularly use AI in at least one business function. But nearly two-thirds say their organizations have not yet begun to scale AI across the enterprise.

Positive indicators of AI’s impact on business

64% say AI has increased the level of innovation in the company. However, only 39% see the impact of AI on financial performance, mostly within individual cases, rather than across the company.

AI-mature companies use AI for transformation, not just efficiency

McKinsey identifies only 6% of companies that have achieved a noticeable effect (>5% EBIT from AI): — 3.6 times more likely to plan transformations — 2.8 times more likely to redesign processes — 3 times more likely to have top-level support — 5 times more likely to scale agents — 4.9 times more likely to invest >20% of digital budget in AI.

Key management practices that distinguish companies successful in implementing AI.

Human-in-the-loop — defined when an AI decision requires human verification.
AI strategy with a clear roadmap .
Active participation of leaders in the implementation of AI.
AI-upskilling — personnel training.
Flexible product teams.
Established processes for rapid model iteration.
A single technology architecture for scaling.
Integration of AI into business processes (not «next to», but inside).
High-quality data products.
Strategic personnel planning taking into account new roles and changes in the work structure.

Efforts to mitigate AI risks are becoming more widespread

51% of companies have already experienced negative impacts from AI, most often due to model inaccuracy (30%). Leaders tend to make mistakes more often, but they are also faster to learn and react.

“I hate artificial intelligence.” Breaking Bad creator calls AI “the world’s most expensive plagiarism machine”

Your own AI on your computer: works without the Internet, free of charge and without limits. Issue 3 of "Show on AI"