Artificial intelligence is increasingly improving workplace productivity, but companies are still struggling to translate those gains into financial performance. As many as 80% of respondents in a McKinsey survey say AI increases their individual productivity, while 37% attribute at least some positive impact on organisational EBIT to the technology. Just 6% belong to the group of so-called AI high performers — organisations that attribute at least 5% of EBIT to AI and consider the value generated by the technology to be significant.
The latest report from QuantumBlack, AI by McKinsey, The state of AI in 2026: On the road to ROI, suggests that companies are entering a new phase of AI transformation.
Deployments now cover more business functions, the use of AI agents is growing, and employees are reporting measurable benefits in their day-to-day work. The main challenge is turning those individual gains into results at the level of the entire organisation.
The experience of companies generating the greatest value from AI suggests that simply adding new tools to existing processes is not enough. The biggest benefits come from fundamentally redesigning how work is organised and executed.
AI is developing faster than organisations
Companies are clearly moving beyond the experimentation stage.
Nearly nine in ten respondents say their organisations now use AI regularly in at least one business function. Around 44% report that AI is already being scaled across the organisation, up from 38% a year earlier.
At the same time, the share of companies using AI in at least three business functions increased from 51% to 56%.
The pace of adoption, however, is uneven.
Among large companies with annual revenues above $1 billion, 54% of respondents say AI is being scaled across the organisation. Among smaller businesses, the figure is closer to one-third.
The gap is even wider when it comes to AI agents.
In large organisations, the share of respondents reporting that AI agents are being scaled increased from 27% to 40% in one year. In smaller companies, the figure remained broadly unchanged at around 22%.
Central Europe has strong AI potential, but a sizeable adoption gap
The global findings are consistent with the picture of Central Europe presented in QuantumBlack’s June report, Central Europe’s AI Opportunity.
According to the analysis, artificial intelligence could generate between EUR 280 billion and EUR 700 billion in economic value annually across the region, equivalent to approximately 6% to 15% of its total net turnover.
At the same time, around 60% of Central Europe’s economy is concentrated in sectors where scaling AI is particularly challenging. These include manufacturing, construction, retail, energy and logistics.
Poland has the largest AI-enabled automation potential in the region, estimated at around EUR 105 billion in additional value by 2030.
Yet the country still faces a significant adoption gap.
Around 29% of Polish consumers use AI privately, while only 8% of companies have implemented the technology at organisational level. By comparison, the average for Western Europe is 28%.
AI agents are changing the build-versus-buy equation
The shift is particularly visible in software development.
Around one in five respondents say their organisations have reached the stage of scaling AI agents that support software engineering and coding. Among large companies, the share rises to 31%.
These new capabilities are also beginning to influence purchasing decisions.
Some 32% of respondents say their organisation has decided not to buy at least one software product or feature because it could build a similar solution internally using AI coding agents.
This trend is most commonly reported by respondents in the technology and healthcare sectors, followed by professional services, energy and natural resources.
Over time, this could shift part of technology budgets away from buying off-the-shelf software and towards building internal capabilities.
Companies are therefore facing a more complex strategic choice: what should be bought, what should be built internally, and where should they develop the capabilities required to integrate and scale solutions created together with external partners.
Higher employee productivity is not enough to improve company performance
The survey shows a clear gap between the impact of AI on individual employees and its effect on the organisation as a whole.
As many as 80% of respondents say AI improves their productivity, while around half say it helps them make better decisions.
These benefits are visible across different levels of organisations.
However, they are not translating into financial performance on the same scale.
Only 37% of respondents attribute at least some positive impact on EBIT to AI — virtually unchanged from a year earlier.
The share of so-called AI high performers, defined as organisations that attribute at least 5% of EBIT to AI and describe the value created by the technology as significant, remains at around 6%.
“One of the clearest findings from this year’s survey is the gap between benefits at the employee level and their impact on overall organisational performance. AI is clearly improving productivity, but companies are not yet seeing a comparable effect on EBIT. The opportunity for leaders is to use employees’ growing AI capabilities to build an operating model that allows end-to-end processes to be redesigned quickly and at scale,” says Michał Miktus, Local Partner at McKinsey & Company in Poland and co-leader of QuantumBlack, AI in Central Europe.
The findings are consistent with the recommendations of Central Europe’s AI Opportunity.
Its authors argue that the largest value does not come from multiplying isolated pilot projects, but from transforming entire business domains — from sales and customer service to manufacturing, supply chains and software development.
The key is to embed AI directly into processes, link deployments to measurable outcomes, and adapt the operating model, skills base and management practices to support scaling.
Leaders redesign work instead of simply adding AI to existing processes
Companies that generate the highest value from AI follow a clear pattern.
Nearly three-quarters of AI high performers are fundamentally redesigning workflows using artificial intelligence. Among other organisations, only around one-quarter are doing the same.
A year earlier, 55% of companies in the high-performing group reported such changes.
The difference is also visible in the level of ambition.
High-performing organisations use AI to improve efficiency just as often as others, but they are far more likely to pursue growth and innovation at the same time.
They are also more than three times as likely to say they expect AI to fundamentally transform their entire business over the next three years.
AI high performers are also more likely to use management practices that help translate technology into business results.
They involve senior leadership, measure the impact of AI initiatives, plan skills strategically, manage costs and control risks.
In addition, they are more than twice as likely to allocate over 15% of their total ICT budget to AI.
“The companies achieving the strongest results stand out primarily because of the consistency of their approach. They combine efficiency goals with growth and innovation, redesign processes rather than layering AI onto existing ways of working, and support implementation with leadership engagement, the right human-in-the-loop model, impact measurement and risk management. Their advantage cannot be reduced to larger budgets alone,” Michał Miktus says.
AI costs are becoming a new management parameter
As AI deployments scale, operating costs are becoming increasingly important.
These include the cost of tokens, computing power and infrastructure required to run AI systems.
One in five respondents says such costs have already limited AI usage in their organisation.
The issue affects companies of different sizes and across multiple sectors.
Despite growing cost pressure, organisations continue to invest heavily in artificial intelligence.
Around 28% of respondents say their companies already spend more than 10% of their total ICT budget on AI technologies, while 60% expect spending to increase further over the next 12 months.
This is changing how companies manage the economics of AI.
Alongside questions about technological capabilities, organisations are increasingly focusing on value measurement, cost control and selecting use cases where spending can produce scalable business results.
Expectations of AI-driven job cuts are still running ahead of reality
The growing use of artificial intelligence is also changing expectations around employment.
Some 39% of respondents expect AI to reduce headcount in their organisations over the next year, while 43% anticipate little or no change.
So far, however, previous expectations have exceeded the actual scale of job reductions.
In McKinsey’s 2025 survey, 32% of respondents expected AI to lead to workforce reductions during the following year.
In the 2026 edition, only 14% reported that AI had actually contributed to a reduction in their organisation’s total headcount over the previous 12 months.
At the same time, only 13% of respondents say AI makes them worried about their own employment prospects.
The findings suggest that expectations about rapid AI-driven workforce contraction remain significantly ahead of what companies are currently experiencing.
The next phase of AI will be about redesigning entire operating models
The report suggests that the organisations most likely to gain an advantage in the next phase of AI transformation will be those that use employees’ growing AI skills to redesign entire operating models.
As artificial intelligence becomes embedded in a growing number of business processes, lasting financial returns will depend not only on adoption itself, but also on the organisation’s ability to change how it operates.
The gap between productivity gains and financial results is therefore becoming one of the central challenges of enterprise AI.
Companies may increasingly find that the question is no longer whether employees can use AI effectively, but whether the organisation can restructure workflows, decision-making, responsibilities and technology infrastructure around those capabilities.
About the research
The report The state of AI in 2026: On the road to ROI is based on a global McKinsey survey conducted online between May 4 and June 8, 2026.
The survey included 1,719 respondents from 97 countries, representing different regions, industries, company sizes, business functions and levels of professional experience.
Some 36% of respondents work in organisations with annual revenues exceeding $1 billion.
The results were weighted to reflect individual countries’ shares of global GDP.
Sources: QuantumBlack, AI by McKinsey, The state of AI in 2026: On the road to ROI; Central Europe’s AI Opportunity; McKinsey & Company.
