Artificial intelligence is moving deeper into the enterprise, but McKinsey’s latest global survey suggests that widespread adoption has not yet translated into equally widespread financial returns.
In its “The State of AI in 2026: On the Road to ROI” survey, McKinsey found that companies are scaling AI across more business functions, deploying AI agents and coding agents, and increasing investment. At the same time, the share of organizations reporting meaningful AI-driven EBIT impact has barely changed from the previous year.
The result is a striking gap: employees are seeing significant productivity gains from AI, while many companies are still struggling to turn those gains into measurable bottom-line results.
McKinsey’s State of AI 2026: The Biggest Findings
McKinsey surveyed 1,719 participants across 97 countries between May 4 and June 8, 2026. The respondents represented different industries, company sizes, functions and organizational roles, with the data weighted according to each country’s contribution to global GDP.
Several findings stand out.
AI adoption is moving beyond experimentation
Nearly nine in ten respondents say their organizations regularly use AI in at least one business function.
More importantly, 44% say AI is now scaling across their enterprise, up from 38% a year earlier.
AI is also spreading across departments. The percentage of respondents saying their organizations use AI in at least three business functions increased from 51% to 56%.
Large organizations are moving faster. Among companies generating at least $1 billion in annual revenue, 54% report scaling AI across the enterprise, compared with about one-third of smaller organizations.
AI Agents Are Becoming a Major Enterprise Trend
The 2026 survey shows that businesses are increasingly moving from traditional generative AI tools toward systems capable of performing tasks with greater autonomy.
Among organizations with more than $1 billion in revenue, 40% report scaling AI agents, compared with 27% in the previous year’s survey.
Smaller organizations have been much slower to make the same transition, with the share reporting scaled agent deployment remaining around 22%.
AI agents are particularly common in:
- IT
- Knowledge management
- Software engineering
- Marketing and sales
- Supply-chain operations
- Manufacturing
McKinsey says technology companies report particularly high use of agents in software engineering, while retailers and consumer companies are more likely to deploy them in marketing and sales. Manufacturers are using them in areas such as supply chain and production.
Coding Agents Are Changing the Build-vs-Buy Equation
One of the more significant findings involves AI-powered software development.
About two in ten respondents say their organizations are already scaling software coding agents, with the figure reaching 31% among larger enterprises.
But the more interesting development is what companies are doing with those tools.
Thirty-two percent of respondents say their organizations decided against buying at least one software product or feature because they could build the functionality internally using agentic coding tools.
That could eventually affect the economics of enterprise software.
Instead of purchasing a specialized application, companies may increasingly decide that an internal AI-powered engineering team can build a customized alternative faster or more cheaply.
This does not mean traditional SaaS is disappearing. But it suggests AI coding agents could gradually change how businesses evaluate software purchases.
80% Say AI Has Improved Their Productivity
Perhaps the clearest evidence of AI’s current value comes from individual employees.
Eight in ten respondents say AI has improved their individual productivity.
Around half also say AI has helped them make better decisions, while many respondents report improvements in skills and other workplace capabilities.
This creates an important distinction.
AI appears to be delivering value at the individual level much faster than at the enterprise financial level.
Employees may finish work faster, generate ideas more quickly or automate repetitive tasks, but companies still have to redesign processes, manage costs and change organizational structures before those gains become substantial financial improvements.
Enterprise AI ROI Remains the Big Problem
This is arguably the most important conclusion from McKinsey’s 2026 research.
Despite the rapid expansion of AI, only 37% of respondents say AI has contributed positively to their organization’s EBIT.
That percentage is essentially unchanged from the previous year.
Even more striking, McKinsey’s group of AI high performers—organizations where AI contributes at least 5% of EBIT and has a significant overall impact—represents only about 6% of respondents.
In other words, companies are using more AI, but the proportion reporting major financial impact has not increased at the same pace.
That is why McKinsey frames the 2026 report around the journey “to ROI.”
AI Costs Are Becoming a Constraint
The economics of AI are becoming another challenge.
About one in five respondents says AI operating costs are constraining their organization’s use of the technology.
Those costs include expenses such as AI model and token usage.
Yet companies are not backing away from AI investment.
Sixty percent of respondents expect their organizations to increase AI investment during the next year, while 28% say AI already accounts for more than 10% of their enterprise-wide information and communications technology budget.
This creates a balancing act for CIOs and business leaders:
More AI → more usage → higher costs → greater pressure to prove ROI.
The Companies Winning With AI Are Doing More Than Adding Chatbots
McKinsey’s research suggests that AI leaders are approaching the technology differently.
Only around 6% of respondents qualify as AI high performers, but these organizations are more likely to use AI for growth and innovation rather than simply cutting costs.
They are also much more willing to redesign existing workflows.
Nearly three-quarters of AI high performers report fundamentally redesigning workflows because of AI, compared with only about one-quarter of other respondents.
That difference may be one of the most important lessons from the report.
Simply putting an AI assistant inside an existing workflow may improve productivity.
But completely redesigning the workflow around what AI can do may create much larger gains.
AI Is Reshaping the Workforce—But Not Yet at the Scale Many Expected
McKinsey’s workforce data also provides an important reality check.
Only 14% of respondents from organizations using AI say AI contributed to an overall decline in workforce size during the past year.
That is considerably lower than the 32% of respondents in last year’s survey who expected workforce reductions over the same period.
However, expectations for the future are becoming more aggressive.
Thirty-nine percent of respondents now expect AI to reduce their organization’s total workforce during the next year.
At the same time, 43% expect little or no AI-related change.
So far, AI-driven workforce reductions have been substantially smaller than companies previously predicted—but expectations are rising again.
Where AI Is Creating Cost and Revenue Gains
The financial impact is not completely absent.
McKinsey found that respondents most frequently report cost reductions from AI in areas such as:
- Supply-chain management
- Service operations
- Manufacturing
Revenue gains are more commonly associated with:
- Marketing and sales
- Product and service development
- Software engineering
This suggests that AI’s economic impact may be highly dependent on where and how it is deployed, rather than simply whether a company has adopted AI.
What McKinsey’s 2026 Report Really Means
The headline takeaway isn’t that AI is failing.
It is almost the opposite.
AI adoption is accelerating, agents are becoming more common, employees are reporting major productivity improvements, companies are increasing investment and organizations are beginning to rebuild software internally using coding agents.
The problem is that enterprise transformation is harder than AI adoption.
A company can give thousands of employees access to an AI assistant in weeks.
Turning that usage into higher revenue, lower operating costs or a structurally different business model requires much more.
It requires workflow redesign, leadership commitment, AI governance, cost management, data infrastructure and measurement.
McKinsey’s highest-performing organizations appear to understand this distinction. They are not simply adding AI tools to existing processes; they are changing how the business operates around AI.
The Bottom Line
McKinsey’s State of AI 2026 paints a picture of an AI market entering its next phase.
The experimentation era is giving way to scaling.
AI agents and coding agents are moving into real business workflows. Individual employees are already seeing substantial productivity gains. Companies are committing more money to AI and, in some cases, replacing purchased software with internally built AI-powered solutions.
But the biggest question remains unanswered for much of the corporate world:
Can productivity gains become measurable enterprise-level financial returns?
McKinsey’s answer so far is that the transition is happening—but slowly.
The organizations most likely to win may not be the ones using the most AI tools. They may be the ones willing to redesign their businesses around what AI makes possible.
FAQ
What is McKinsey’s State of AI 2026 report?
It is McKinsey’s latest global survey examining how organizations are adopting, scaling and investing in artificial intelligence, including generative AI, AI agents and coding agents.
How many organizations are scaling AI in 2026?
McKinsey found that 44% of respondents report AI scaling across their enterprise, compared with 38% a year earlier.
How many employees say AI improves productivity?
Eight in ten respondents say AI has improved their individual productivity.
Is AI generating significant enterprise ROI?
Not yet for most organizations. About 37% of respondents say AI has contributed positively to organizational EBIT, essentially unchanged from the previous year.
How many companies are scaling AI agents?
Forty percent of respondents from organizations with more than $1 billion in annual revenue report scaling AI agents, compared with 27% the previous year.
Are companies using AI to build their own software?
Yes. Thirty-two percent of respondents say their organizations decided against purchasing at least one software product or feature because they could build it internally using agentic coding tools.
Is AI reducing jobs in 2026?
The workforce impact remains mixed. Fourteen percent of respondents from organizations using AI reported AI-related overall workforce declines during the past year, while 39% expect AI to reduce total employment during the coming year.
What is the biggest lesson from McKinsey’s 2026 AI report?
The biggest lesson is that adopting AI tools is not enough. Organizations achieving the strongest results are redesigning workflows and business processes around AI rather than simply adding AI to existing systems.
