McKinsey: Companies Report AI Productivity Gains, but Profits Still Lag

McKinsey has published the latest edition of its recurring State of AI survey, and the results reveal a gap that consulting firms have been observing for years without quite explaining it. Employees feel that artificial intelligence genuinely makes their work easier. Boards, for the most part, still don't see that effect show up in the financial results.

What the survey found

McKinsey has been surveying managers and employees about AI use at companies on a recurring basis for several years, and this year's edition covered 1,719 respondents from 97 countries. The survey shows that AI adoption is rising almost everywhere, while the number of organizations translating that adoption into measurable profit remains stubbornly low.

Among companies earning more than a billion dollars a year, 40 percent are already scaling AI agents, meaning systems capable of independently carrying out multistep tasks, compared with 27 percent a year earlier. Nearly a third of organizations have chosen to forgo buying off-the-shelf software in favor of building functionality themselves with AI coding tools.

The gap between employee and company

The report's authors call this phenomenon the gap between individual benefits and enterprise-wide impact. An employee may work faster thanks to a chatbot or a coding assistant, but that time saved rarely translates into a countable rise in the whole company's operating profit, because the effects get diluted across processes that AI touches only partially.

A striking finding this year is the gap between individual gains and enterprise impact - Dan Tinkoff, senior partner, McKinsey

Yet only 37 percent of organizations report any positive contribution to EBIT, virtually unchanged from last year - Dan Tinkoff, senior partner, McKinsey

Costs rising faster than returns

The report also points to mounting cost pressure. About 20 percent of respondents admitted that operating expenses tied to maintaining AI systems are already limiting the scale of their use. At the same time, 28 percent of organizations are putting more than 10 percent of their IT budget toward AI tools, and 60 percent plan to increase that investment next year despite uncertainty about the return.

McKinsey also notes that only 6 percent of surveyed companies belong to the group it calls AI high performers, meaning organizations that both attribute at least 5 percent of EBIT to AI and rate its impact as significant. That is a narrower and more demanding definition of success than simply reporting any positive effect at all.

Employment worries

The survey also touched on the labor market. 39 percent of respondents expect their employer to cut jobs because of AI in the coming year, up from 32 percent a year earlier. At the same time, 43 percent do not expect significant changes in overall headcount, and the actual job cuts seen in 2025 turned out, according to earlier data, to be far smaller than previous forecasts had suggested.

For Polish companies these figures carry practical weight, since domestic reports show a similar pattern: employees and managers report a noticeable productivity boost from AI tools, but chief financial officers still struggle to show a hard return on investment in the numbers. McKinsey suggests the problem lies not in the technology itself but in how it is deployed, that is, whether a company is actually rebuilding its processes around AI or merely bolting a tool onto its existing way of working.

The report's authors are not arguing that AI doesn't work. Rather, they point out that the financial payoff requires time and deeper organizational change than simply handing employees a chatbot license. The next edition of the survey will show whether the share of companies crediting AI with a real impact on profits finally starts to climb, or stays frozen at the level of the past two years.