Artificial intelligence is no longer an exception in the business world; it has become the norm. According to McKinsey’s global research, 88% of companies utilize AI regularly in at least one business unit. However, this ubiquity does not equate to depth. Most organizations remain in the trial and pilot phases; only 7% have successfully scaled AI across the entire organization.
The size gap is also significant. Half of the companies with annual revenues exceeding $5 billion have reached the scaling phase, whereas this rate drops to 29% for companies with revenues under $100 million. The most common use cases are knowledge management (40%), marketing and sales (39%), and IT (34%).
Innovation is Rising, Financial Returns Are Limited There is broad consensus that AI accelerates innovation. 64% of respondents state that the technology increases innovation capacity, while 45% report improvements in customer satisfaction and competitive advantage.
However, financial impacts remain limited. Only 39% of respondents report that AI has had an impact on company-wide profit (EBIT). Most within this group state that the impact remains below 5%.
When viewed by function, cost reductions are most prominent in software engineering (56%), manufacturing (56%), and IT (54%). Revenue increases, on the other hand, stand out in sales and marketing (67%), strategy and corporate finance (65%), and product development (62%).
The New Wave: Autonomous AI Agents 62% of companies are experimenting with autonomous AI agents. These systems can plan and execute multi-step tasks independently. However, scaled usage is still rare; across any given business unit, no more than 10% of companies have deployed these technologies at scale.
AI agents are most common in IT (9%), knowledge management (8%), and marketing/sales (7%). The technology, media, telecommunications, and healthcare sectors are leading the charge.
The "High Performers" in the Minority The report identifies only 6% of respondents as "AI high performers." This group comprises organizations reporting AI-driven earnings equivalent to at least 5% of company profit.
These companies position AI not merely for efficiency, but as a vehicle for transformation. They are 3.6 times more likely to set transformative goals than others. They view growth (82%) and innovation (79%) as primary objectives, rather than just cost-cutting.
Furthermore, instead of simply adding AI to existing processes, these companies redesign their workflows. The rate of those doing so is 57%—almost three times the general average (20%).
Leadership commitment also makes a distinct difference: nearly half of the high-performing organizations report that their top management takes ownership of AI initiatives.
Financially, they are bolder as well. 35% of these institutions allocate more than one-fifth of their total digital budget to AI; this rate is only 7% for others.
Risk Management is Maturing The report indicates that companies are becoming more mature in managing AI risks. On average, an organization actively monitors four different risks. The most frequently encountered issue is "accuracy" errors (hallucinations) at 30%, which is also the most mitigated risk (54%).
Cybersecurity comes in second (51%). Lack of explainability, while less addressed, remains a significant issue (experienced by 14%, mitigated by 28%).
High-performing companies encounter more risks but are also more successful at managing them. The report calls this the "high-performance paradox": those who use AI more deeply and aggressively also face a higher risk of error.
AI in 2025: By the Numbers
- 88% of companies use AI in at least one area.
- 62% are experimenting with AI agents.
- 39% reported an EBIT impact.
- 64% experienced an increase in innovation.
- 56% reduced manufacturing costs.
- 32% expect a decrease in employment next year.
- 51% are managing cybersecurity risks.
For more information, you can review the McKinsey Report.