McKinsey's newest State of AI survey delivers a blunt verdict on enterprise AI spending: productivity feels real, but profit mostly does not. Matt Britton, an AI transformation speaker who advises Fortune 500 leaders on moving AI from pilot to production, argues this gap is not a technology problem at all. It is an operating-model problem, and the data now proves it at scale.
According to the 2026 survey, the percentage of respondents who say AI has had a positive impact on their company's overall EBIT is 37%, almost unchanged from the previous year, while only about 6% of the total are "AI High Performers" who have boosted EBIT by 5% or more through AI and achieved significant impact, even though 80% of individuals report improved productivity and about 50% report improved decision-making . That is the headline every executive team needs to sit with. Four out of five employees feel the lift from AI tools, yet barely one in sixteen organizations can show it on the income statement.
Britton has spent the past two years telling boardrooms that this exact gap would widen before it closed, and McKinsey's numbers confirm it. Under a stricter definition, just 6 percent of organizations qualify as AI high performers, attributing at least 5 percent of EBIT to AI and calling the impact significant, while 63 percent report no measurable enterprise earnings impact at all . This article unpacks why that divide exists, what the 6% are doing differently, and the operating changes that move AI pilots into durable production value. It draws on Britton's keynote research as founder of FutureProof, his AI transformation advisory, and the arguments laid out in his bestselling book Generation AI.
What McKinsey's State of AI 2026 Survey Reveals About Enterprise AI ROI
McKinsey surveyed more than 1,700 leaders across nearly 100 countries in 2026, and the headline finding is a stubborn flat line. Enterprise AI ROI has not moved meaningfully in twelve months despite record spending on tools and talent. Britton frames this as the defining business story of the year: adoption curves keep climbing while value curves stay flat.
The productivity-to-profit gap is the clearest signal yet that tool deployment alone does not create enterprise value. Workers report real personal gains from AI assistants, chatbots, and copilots embedded in daily tasks. But individual time savings rarely translate into organizational EBIT unless the surrounding workflow, staffing model, and decision rights change with them. Britton calls this the difference between automating a task and redesigning the work around it, a distinction he unpacks in his intelligence enablement keynote.
The implication for CFOs and boards is direct. Measuring AI success by usage metrics, like licenses purchased or prompts run, misses the only metric that matters to the business: earnings impact. Going forward, Britton argues that every enterprise AI initiative should carry an EBIT-linked hypothesis from day one, not a retrofit measurement process added after the pilot already launched.
Inside the 6% Club: What AI High Performers Do Differently
McKinsey's "AI high performers" are a narrow, well-defined group. McKinsey defines AI high performers as respondents who attribute at least 5% of EBIT to AI and describe its value as significant, and they make up about 6% of respondents, unchanged from 2025 . That stability across two survey cycles is itself notable. It means the 6% are not early movers who will simply be joined by the rest of the market as tools mature; they represent a structurally different approach to deployment.
The behavioral gap shows up clearly in budget allocation and workflow commitment. High performers are more than twice as likely as others to spend over 15 percent of their ICT budget on AI . More tellingly, the redesign data is stark: McKinsey places the most emphasis on the redesign of business processes to explain this gap, and when comparing 25 organizational initiatives, the one most strongly associated with EBIT improvement through generative AI was fundamentally rebuilding workflows with AI as the premise, yet only 21% of companies using generative AI have actually fundamentally redesigned any of their business processes .
Britton's reading of this data is unambiguous: tool selection is not the variable separating winners from the other 94%. The 6% treated AI as a reason to rebuild how work flows through their organization, not as a plugin for the workflow that already existed. That single choice, made at the leadership level rather than the IT level, explains most of the gap McKinsey measured.
- Budget commitment: High performers invest more than double the AI share of ICT spend compared to peers.
- Workflow redesign: Only 21% of all AI-using companies have fundamentally rebuilt a business process around AI, yet this is the single strongest predictor of EBIT gains.
- Leadership ownership: High performers report stronger executive accountability for AI outcomes rather than delegated budget approval alone.
- Measurement discipline: Most non-high-performers still lack a defined process to quantify AI's financial impact.
Why Most Enterprise AI Pilots Stall Before Production
The pattern McKinsey documents in 2026 echoes what Britton has observed across more than 500 keynotes delivered to enterprise audiences on five continents. Pilots launch with enthusiasm, generate strong demo results, and then stall the moment they meet the existing operating model. The technology works in isolation. The organization around it was never built to absorb the change.
McKinsey's own organizational research names this constraint explicitly. Recent McKinsey AI readiness research reinforces that gap, and as AI capabilities advance, the existing organization can become the constraint . In practice, this means the reporting lines, approval chains, and job descriptions built for a pre-AI workforce actively throttle the value a new capability could otherwise unlock. A faster model cannot outrun a slower org chart.
The scale of the required change is larger than most leadership teams assume. Separate McKinsey organizational research puts a number on the exposure: 75% of roles need fundamental reshaping right now to operate compatibly with agentic AI tools already in deployment. That is not a future-state statistic. It describes the condition most enterprises are already operating under without having acted on it.
Britton's forward view is that this gap will not close through better prompting or newer models. It will close only when leadership teams treat workflow redesign, role redesign, and decision-rights redesign as part of the AI rollout itself, not a follow-up project scheduled for next year. Enterprises that wait for a "mature enough" model before redesigning the operating model will still be in the 63% a year from now.
Workflow Redesign vs. Bolt-On AI: The Operating Model Difference
The distinction between bolting AI onto existing work and rebuilding the work around AI sounds subtle until it is quantified. McKinsey's organizational research is explicit about which approach compounds value over time. Workflows change as teams learn what AI makes possible, rather than simply adding AI to existing work , and this is the dividing line between pilots that scale and pilots that quietly die after the demo.
Bolt-on AI preserves the org chart, the approval chain, and the job descriptions exactly as they were, then inserts a model somewhere inside them. It produces measurable individual productivity, which explains McKinsey's 80% figure, but it rarely changes headcount, cycle time, or cost structure enough to move EBIT. Workflow redesign instead starts from the outcome a team is trying to produce and rebuilds roles, handoffs, and review steps around what AI and humans can now do together.
McKinsey's broader organizational guidance for 2026 reinforces this as a leadership mandate rather than a technical one. Leaders can capture the larger opportunity from AI by organizing work around outcomes, reshaping the talent system to reflect the changing mix of human and AI work, building the mechanisms that allow change to compound, and putting clear authority and accountability behind the transformation . Britton frames this as the real content of an enterprise AI strategy. Everything else, including model choice and vendor selection, is implementation detail layered on top of that decision.
The forward view for 2027 planning cycles is that workflow redesign moves from a nice-to-have to a board-level requirement. Companies that keep funding pilots without funding the organizational redesign around them will keep producing McKinsey's flat 37% line. Those that commit leadership time to outcome-based redesign are the only group with a plausible path into the 6% club.
The Org-Design Playbook for Scaling AI From Pilot to Production
Britton's keynote framework for enterprise audiences breaks the pilot-to-production transition into four leadership moves, each grounded in the McKinsey data above. First, leaders must define the EBIT hypothesis before launch, not after. Second, they must assign a single executive owner for outcomes rather than distributing accountability across IT, operations, and a steering committee that never quite converges.
Third, leaders must be willing to eliminate or restructure roles and approval steps that the AI workflow makes redundant, something far fewer than the needed share of companies have done so far based on the 21% redesign figure cited above. Fourth, measurement has to be built in from day one. Many organizations have no defined process to quantify AI impact, with only about 20% of non-high-performers having one in place , which means most companies cannot even tell whether their pilots are working.
Britton also points to real-time organizational listening as an underused lever in this redesign process. Through his work as Executive Chairman of Suzy, he has seen how continuous feedback loops, applied internally to employees and externally to customers, help leadership teams spot where a redesigned workflow is actually landing versus where it is quietly being bypassed. That kind of signal is what separates a redesign announced in a town hall from a redesign that sticks six months later.
The forward view here is straightforward: the 6% will not stay at 6% indefinitely, but the companies that join them will be the ones that treated 2026 and 2027 as the years to rebuild operating models, not the years to buy more licenses. Britton's argument to Fortune 500 audiences is that the next eighteen months are the window to move, because the organizational gap only widens as agentic AI capabilities keep advancing faster than most org charts can absorb them.
Key Takeaways for Business Leaders
- Audit every active AI pilot against a single question: has the workflow been rebuilt, or has AI simply been added to the old one?
- Assign one accountable executive owner per AI initiative with EBIT impact tied directly to their role, not a shared steering committee.
- Build a measurement process before launch, since most non-high-performers still lack any defined way to quantify AI's financial return.
- Prioritize role and process redesign over additional tool procurement, given that workflow rebuilding is the strongest predictor of EBIT gains in the 2026 data.
- Invest leadership time in outcome-based redesign now, since the gap between AI capability and organizational readiness is widening, not closing.
Frequently Asked Questions
What is McKinsey's "6% club" in AI adoption?
McKinsey's 2026 State of AI survey found that only about 6% of organizations qualify as "AI high performers," meaning they attribute at least 5% of EBIT to AI and describe the value as significant. The figure has stayed flat since 2025, even as broader AI adoption and spending have continued to rise across nearly every industry.
Why do most enterprise AI pilots fail to reach production?
Most pilots stall because leaders bolt AI onto an unchanged workflow instead of redesigning the roles, approvals, and decision rights around it. McKinsey's data shows that fundamentally rebuilding workflows, not model selection, is the strongest predictor of EBIT impact, yet only about 21% of AI-using companies have actually done this.
What does an AI transformation speaker actually help leadership teams do?
An AI transformation speaker like Matt Britton helps executive teams move past tool-focused thinking toward operating-model change. His keynotes connect data like McKinsey's State of AI findings to concrete org-design decisions, including role redesign, accountability structures, and measurement systems that convert AI pilots into EBIT impact.
How can a company measure true AI ROI instead of just productivity?
True AI ROI requires tracking EBIT attribution, not just usage or sentiment metrics like "feels more productive." Leaders should define a financial hypothesis before launching a pilot, assign executive ownership for outcomes, and build a measurement process from day one, since most organizations currently lack any formal way to quantify AI's bottom-line effect.
The McKinsey data is unambiguous: the gap between AI adoption and AI value is an organizational choice, not a technology ceiling. Matt Britton has built his keynote platform around exactly this argument, helping Fortune 500 leadership teams translate research like this into the operating-model decisions that move companies from pilot purgatory into the 6% club. His talks connect directly to the frameworks in Generation AI and the ongoing conversation on The Speed of Culture podcast at speedofculture.co.
Leadership teams preparing 2027 AI budgets should treat this moment as the window to act, before the organizational gap widens further. To bring these insights to your next event, explore Matt Britton's AI keynote speaker platform or visit Speaker HQ to contact his team directly.



