🔍 Read the full analysis: Most Companies Use AI, But Few Are Seeing Profits on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
AI adoption is spreading, but major surveys suggest few companies can tie it to significant financial returns. McKinsey found 6% of respondents’ organisations qualified as high performers, while BCG found 5% of firms were getting value from AI at scale. The surveys point to workflow redesign, data readiness, skills and clear controls as factors shaping results.
Most companies now use AI somewhere, but only a small share report substantial financial returns, according to surveys reviewed by Thorsten Meyer. McKinsey’s 2026 survey found that 6% of respondents’ organisations met its definition of high performers, attributing at least 5% of earnings before interest and taxes (EBIT) to AI.
The figures point to a gap between adoption and reported business value. In the McKinsey survey of 1,719 respondents, 37% said AI had some effect on their organisation’s EBIT, a share reported as roughly unchanged from the previous year. The survey’s high-performer category was narrower: organisations had to attribute at least 5% of EBIT to AI. The findings were reported by The Register and TechTimes.
Boston Consulting Group’s survey of 1,250 companies showed a similar divide. BCG said 5% of firms were getting value from AI at scale, while 60% reported no material value. These are survey findings, not audited financial results, and the measures used by McKinsey and BCG are not identical.
AI use is also uneven by company size and location. Eurostat reported that 20.0% of EU enterprises with at least 10 employees used AI in 2025, up from 13.5% in 2024. The figure was 55% among large enterprises and 17% among small ones. The European statistics measure reported use, not whether AI improved profits.
Companies are also experimenting with AI agents, systems designed to take steps within set limits rather than only answer prompts. BCG respondents attributed about 17% of AI value to agents and expected the share to reach 29% by 2028. McKinsey’s survey, as reported by The Register and TechTimes, found that 40% of companies with more than $1 billion in revenue said they were scaling agents, up from 27%.
Infographic 01 · Who gets value from AI
Most companies use AI. About 6 percent profit from it in a big way
Adoption keeps rising. The share of companies that can show a large effect on profit has barely moved.
Share of McKinsey’s 2026 respondents who attribute at least 5 percent of their company’s EBIT to AI and call the impact significant. Flat versus 2025, as reported by The Register.
Wide use, thin profit
Different surveys, different bases: read each line with its source
Sources: McKinsey State of AI 2026 (n = 1,719), as reported by The Register, 25 Aug 2026 and TechTimes, 26 Aug 2026 · BCG, The Widening AI Value Gap, Sept 2025 (n = 1,250) · Eurostat, 11 Dec 2025 · Eurostat Statistics Explained (enterprises with 10+ employees). Survey figures are self-reported by executives; Eurostat counts use, not value.
Powered by Thorsten Meyer AI
Why AI Adoption Has Not Paid Off
The gap between companies trying AI and companies reporting significant returns matters because adoption figures alone can give an incomplete picture of business impact. A firm may use AI in a limited trial or a single team without changing its operating costs, output or revenue enough to affect company-wide results. The survey findings suggest that use is not the same as measurable profit.
The reports point toward changes to entire workflows, rather than adding an AI tool to an unchanged process, as a way companies seek larger returns. That can involve software handling routine steps, checking records and preparing responses, while employees supervise exceptions and remain responsible for outcomes. Such changes may affect how companies organise work and where staff time is spent. They also require decisions about permissions, review and accountability.
For workers, productivity gains and concerns about job reductions appear side by side in the McKinsey results. Among respondents who use AI, 80% said it made them more productive. At the same time, 39% expected job cuts, up from 32% in the prior survey. Those are respondents’ reported experiences and expectations; they do not establish how many jobs will be eliminated.
Products Worth Considering
Apple MacBook Pro 16-inch with M1 Pro, 10-core CPU, 16-core GPU, 16GB RAM, 1TB SSD, Silver
Apple MacBook Pro MPHF3LL/A 14.2-Inch Laptop | M2 Pro 12-Core Chip 16GB Unified Memory 1TB SSD | 19-Core GPU Liquid Retina XDR Display Space
Logitech MX Anywhere 3S for Mac Compact Wireless Mouse, Space Gray
As an affiliate, we earn on qualifying purchases.
From Digitised Tasks to AI Workflows
Earlier digital transformation efforts often moved paper records and existing steps into software while leaving people to interpret cases and make decisions. AI systems can work with less-structured material, such as emails, scanned invoices and support tickets, and can retrieve information from other systems or take permitted actions. The shift described in the source material is from digitising a process to letting software handle parts of the work.
That does not mean every task should be delegated to an agent. The source attributes to Gartner guidance that agents are suited to situations requiring decisions, conventional automation to routine workflows, and assistants to simple information retrieval. It also reports Gartner forecasts that more than 40% of agentic AI projects could be cancelled by the end of 2027 because of cost, unclear value or inadequate risk controls. A separate Gartner forecast says organisations will abandon 60% of AI projects that lack AI-ready data.
Barriers also include skills. Eurostat data cited in the source found that among EU firms that considered AI but did not use it, 71% named a lack of expertise. BCG’s survey also showed a difference between leaders and laggards: leaders had deployed 62% of their AI initiatives, compared with 12% for laggards. These figures describe survey responses and deployment rates; they do not by themselves establish that deployment caused stronger financial performance.
What the Survey Figures Cannot Show
The figures do not provide a single, directly comparable measure of AI’s financial effect. Eurostat tracks whether enterprises use AI, while McKinsey and BCG ask respondents about business impact and value using their own survey definitions. The results are self-reported survey findings, not a common audit of company accounts, and they cover different groups and measures.
It is also unclear from the reported findings which specific AI tools or applications produced the profit effects attributed to AI, how those effects were calculated, or how much came from cost savings versus new revenue. The surveys do not establish that AI alone caused reported changes in EBIT. The source material also does not give the full methodology or question wording for every cited result, limiting comparisons across studies.
Forecasts about cancelled projects and expected agent adoption describe projections, not final outcomes. The cited figures do not establish how many projects will be cancelled, whether firms’ expectations for 2028 will be met, or how reported productivity gains will affect employment over time.
Tracking Returns Beyond Adoption
The next test for companies is whether AI projects move from trials into sustained use with results that can be measured against a clear baseline. That means tracking the cost of implementation and oversight alongside changes in processing time, error rates, service levels or financial performance. Companies will also need to identify who is accountable when an AI system takes an action or produces an incorrect result.
Gartner’s cited forecasts put project cancellations and data readiness on the watch list through the end of 2027. BCG’s survey points to continued growth in the share of AI value attributed to agents by 2028, but that figure is an expectation, not a confirmed result. Future survey releases and company disclosures may clarify whether adoption is producing broader, measurable returns.
For now, the available evidence supports a measured conclusion: AI use is expanding, but surveys do not show substantial profit gains for most companies. The scale of returns, their causes and their effects on workers remain developing questions.
Key Questions
How many companies are seeing substantial profit gains from AI?
In McKinsey’s 2026 survey, 6% of respondents’ organisations qualified as high performers by attributing at least 5% of EBIT to AI. BCG separately reported that 5% of surveyed companies were getting value from AI at scale. The studies use different measures, so the figures should not be treated as directly equivalent.
Does AI use mean a company is making money from it?
No. Adoption and financial impact are different measures. Eurostat reported AI use at 20.0% of EU enterprises with at least 10 employees in 2025, while separate company surveys found that only a small share reported substantial or scaled value.
Why do some AI projects fail to deliver value?
The source material cites cost, unclear business value, weak risk controls, integration difficulties and a lack of AI-ready data as possible reasons projects stall or are abandoned. In Eurostat data cited in the source, 71% of EU firms that considered AI but did not use it named missing expertise.
Are AI agents replacing workers?
The cited surveys do not establish how many jobs AI agents will replace. McKinsey respondents who use AI reported productivity gains, while a rising share expected job cuts. Those expectations are not confirmed employment outcomes.
When could more evidence on AI returns be available?
The source cites Gartner forecasts about project cancellations through the end of 2027 and BCG respondents’ expectations for agent-related value by 2028. Those are projections; future company disclosures and survey results will show whether they materialise.
Source: ThorstenMeyerAI.com
