McKinsey's study The state of AI in 2026: On the road to ROI, published on August 25, 2026, finds that most companies cannot show material EBIT impact from their AI investments despite rising spend. The report argues that simply deploying AI or increasing individual worker productivity does not automatically translate into measurable financial returns.
What sets leading adopters apart?
According to the study, three‑quarters of the companies that are successful with AI have fundamentally redesigned their workflows. By contrast, only one quarter of other organizations have made comparable changes. McKinsey concludes that ROI depends less on how much a company spends on AI and more on how ready it is to radically reconfigure processes and operations around AI capabilities.
Key findings in numbers
- 80% of respondents say AI improves individual productivity, but only 37% reported a positive EBIT impact.
- Only 6% of organizations consider themselves leading AI users; these leaders attribute at least 5% of their EBIT to AI.
- Leaders are more than twice as likely to allocate over 15% of their ICT budget to AI.
- Currently, 28% of organizations devote more than 10% of their total ICT budget to AI, and 60% expect investments to rise over the next 12 months.
- One in five respondents says AI operating costs — including token costs — are already a constraint that limits usage.
- 32% of respondents said their organization decided not to purchase at least one software product or feature because in‑house AI development tools could replace it.
Adoption breadth and depth
- Nearly nine in ten respondents regularly use AI in at least one business area.
- 44% of respondents say AI is widely adopted at the enterprise level, up from 38% last year.
- The share of organizations using AI in at least three areas rose from 51% to 56%.
- Acceleration in AI use is concentrated among large companies: among firms with more than $1 billion in annual revenue, 54% report enterprise‑wide AI use; for smaller firms the rate is about one third.
- The proportion of large‑company respondents who say their organization widely uses AI agents rose from 27% to 40% year over year, while smaller organizations remained around 22%.
- About one in five respondents says their organization widely uses coding‑assistance AI agents; among large firms that share reaches 31%.
Workforce impacts and concerns
- 39% of respondents expect headcount at their organization to decline over the next year because of AI, while 43% expect little or no change.
- Only 14% believed AI played a role in headcount reductions at their organization over the past 12 months.
- Just 13% of respondents said the technology makes them worried about their own career prospects.
Hungarian perspective: Prompt Magyarország and economic potential
McKinsey's June report Prompt Magyarország: a mesterséges intelligencia hatása a gazdaság versenyképességére reached a similar conclusion. It estimates that, through AI‑enabled automation, roughly €15 billion could be generated in Hungary by 2030 — equivalent to about 6–7% of current GDP. The report stresses that this figure represents the economic potential from partially or fully automatable activities, and is not a direct GDP‑growth projection.
Conclusion
The McKinsey analysis suggests that extracting measurable financial returns from AI hinges on organizational readiness and the extent of process redesign. While AI can raise individual productivity, only companies that redesign workflows and integrate AI strategically are likely to convert that productivity into meaningful EBIT impact.
