AI is everywhere in organizations. Employees are using it. Leaders are investing in it. Adoption continues to climb.
And yet the business impact hasn’t budged nearly as much.
McKinsey’s new State of AI in 2026: On the Road to ROI survey of more than 1,700 respondents across 97 countries reveals a contradiction that has been building for the past few years.
AI adoption is growing. So why isn’t the value growing with it?
McKinsey’s research points to an important distinction between organizations that use AI and those that are creating significant value from it: High performers aren’t simply adopting more AI. They are changing their businesses around it.
Compared with other organizations, AI high performers are 3.3 times more likely to aim to fundamentally transform their business with AI. Nearly three-quarters report fundamentally redesigning workflows because of AI, compared with only one-quarter of other respondents.
They are also more likely to pursue growth and innovation in addition to efficiency, actively manage AI risks, measure the impact of AI initiatives, and have senior leaders who demonstrate commitment to those initiatives.
In other words, the organizations getting the greatest returns aren’t asking only, “Where can we use AI?” They’re asking a much harder question: “What needs to change for AI to create meaningful value here?”
One of the most interesting findings in McKinsey’s survey may be the difference between individual and organizational results.
People are clearly benefiting from AI. They’re saving time, making decisions and developing new skills.
That matters.
But organizations don’t create value simply because individual employees save time. Value comes from what happens with that time: whether work moves faster, decisions improve, customers receive better service, innovation accelerates or costs decline.
That requires more than individual productivity. It requires the ability to capture what employees learn, share knowledge, adapt processes and develop new capabilities as the work changes.
That’s why a learning culture isn’t just about training anymore. It’s increasingly about creating an organization that can learn and adapt continuously.
And that’s where many AI initiatives still have work to do.
For the past several years, organizations have understandably focused on acquiring AI capabilities, experimenting with tools and identifying promising use cases.
The next phase requires more attention to the environment in which those tools operate.
If policies are outdated, procedures contradict one another, critical information lives in employees’ heads, or no one knows which version of a document is current, AI doesn’t make those problems disappear. It can amplify them.
We recently described the kind of employee AI increasingly rewards: someone who documents decisions, follows consistent processes and makes knowledge easier for others to use. The larger lesson behind AI Has a New Favorite Employee applies at the organizational level, too. AI performs better when the information surrounding it is reliable, structured and current.
Organizations need knowledge that both people and AI systems can use. That means treating documentation and knowledge management as part of AI infrastructure, not administrative housekeeping.
Teaching people how to use a particular AI tool is only one piece of the challenge. Employees need to know how to evaluate AI-generated information, recognize when something looks wrong, ask better questions and apply their own expertise and judgment.
As we explored in The Most Important AI Skill Isn’t Prompting, It’s Questioning, employees who work effectively with AI need more than technical proficiency. They need the confidence and context to challenge an answer rather than simply accept it.
They also need opportunities to develop those capabilities as the technology changes.
Interestingly, about half of McKinsey’s respondents say AI has already helped them develop new skills. The opportunity now is to turn that informal learning into organizational capability.
Organizations can automate unnecessary steps, accelerate confusing workflows and make bad processes happen faster. The organizations seeing the greatest value are doing something different: redesigning workflows around what AI makes possible.
That starts by examining the process itself:
We’ve made the case before that your processes may be too complicated. AI makes simplification even more important. Automating complexity is still complexity.
Sometimes the smartest AI strategy starts by simplifying the work before automating it.
Roles may shift. Decision-making may change. Employees may need new skills. Managers may need to supervise work performed partly by people and partly by AI. Leaders need to establish expectations, guardrails and measures of success.
That’s also why managing AI isn’t managing software. As AI moves from simply executing commands to generating recommendations and performing increasingly complex work, leaders have to think differently about objectives, oversight, judgment and accountability.
Technology implementation may be the visible part of AI transformation. Organizational change is the harder part.
McKinsey’s findings also provide an interesting reality check on one of the biggest AI conversations: workforce reduction.
In last year’s survey, 32% of respondents expected AI to decrease their organization’s workforce during the coming year. Only 14% now report that AI actually contributed to an overall workforce decline during that period. Two-thirds report little or no AI-related change in total employment.
That doesn’t mean workforce disruption isn’t coming. In fact, expectations have risen again: 39% now anticipate AI-related workforce declines during the next year. But so far, predictions about workforce reduction are running ahead of organizations’ actual experience.
The more immediate challenge may be less about eliminating work than changing how work gets done. That makes reskilling, knowledge transfer, process design and change leadership even more important.
There is plenty of encouraging news in McKinsey’s latest survey:
The technology is moving. Now organizations have to move with it.
The AI adoption problem is increasingly being solved. The AI value problem isn’t. Closing that gap requires organizations to look beyond the tools themselves and strengthen the knowledge, learning, processes, governance and leadership surrounding them.
For organizations trying to close the gap between AI investment and measurable value, the challenge often isn’t identifying another tool. It’s identifying what needs to change around the technology.
MATC helps organizations strengthen those foundations. We capture and structure critical knowledge, improve documentation, simplify processes, build learning experiences, and help employees and leaders adapt to new ways of working.
We also help organizations identify where AI can genuinely improve performance—and where the knowledge, processes or people supporting it need attention first.
Because adopting AI is becoming easier. Turning it into sustainable business value is where the real work begins.
Value Score vs. Feasibility Score: Why Great AI Ideas Still Fail in Production
The Hidden Cost of Tribal Knowledge Isn’t What You Think
AI Adoption Starts with Trust, Not Training
“The state of AI in 2026: On the road to ROI.” McKinsey. 8/25/26. Accessed 8/31/26. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai