Skip to content

Executive Summary

Almost every large organization is now spending on AI. Far fewer are getting value back. The gap between those two facts is the entire problem this bundle exists to close — and closing it is a leadership problem before it is a technical one. This page is the one-page synthesis: the size of the gap, the model for thinking about the whole effort, why most programs fall into it, and what leadership should do in the first 90 days. The detail lives in the eight tracks and the program-execution layer; this is the map you read first.

The Transformation Problem in Numbers

Adoption is nearly universal and value is not. 89% of organizations now use AI in at least one business function, yet only 37% report any enterprise-level EBIT impact from it, and just ~6% — McKinsey's "AI high performers" — attribute 5% or more of EBIT to AI. Those are executives describing their own organizations in a survey run by a firm that sells AI transformation work. Official statistics count differently and land lower. In May 2026, 19.8% of US firms reported using AI (US Census Bureau, 2026). In 2025, 20.0% of EU enterprises with 10 or more employees did (Eurostat, 2025). Those count firms one vote each, so small businesses dominate. The same survey's AI supplement put the firm share at 18% for November 2025 to January 2026. Weighting that by employment lifts it to 32% (Census CES, 2026). Among very large firms in the Information, Professional Services and Finance sectors, 50-60% used AI, or 60-70% employment-weighted (Census CES, 2026). Those figures come from a Census working paper, meaning researchers' analysis rather than a Bureau release. The consultancy and official figures measure different populations. They are not disagreeing. The pipeline leaks at every stage: Gartner projected that at least 30% of generative-AI projects would be abandoned after proof of concept by the end of 2025 (Gartner, 2024); MIT's study of enterprise deployments found that roughly 95% of gen-AI pilots delivered no measurable P&L impact (MIT NANDA, 2025); and BCG reports that only about 5% of companies are achieving AI value at scale while around 60% capture little or no value (BCG, 2025).

The shortfall is not mainly a technology shortfall. Only about 13% of organizations are the most AI-ready "Pacesetters", a share that has held steady for three years (Cisco, 2025). Only 1% of leaders describe their AI deployment as mature (McKinsey, 2025). BCG's "future-built" companies show about 1.7× the revenue growth and 1.6× the EBIT margins of the 60% it calls stagnating or emerging (BCG, 2025). They also expect 1.4× greater cost reductions than laggards where they apply AI (BCG, 2025). BCG sells transformation work, and the comparison shows an association, not what caused the gap.

The Eight-Track Model

AI transformation is not one program — it is eight interdependent workstreams that succeed or fail together. Most organizations run whichever subset their loudest function owns, and then stall for reasons that live in a track nobody resourced. The framework architecture lays out all eight and how they depend on one another; they group into four layers:

The single most important property of the model is that a track is only as effective as its weakest upstream dependency. Funding Execution on top of unready Foundation produces motion without value; buying a platform without building Talent produces logins without usage. This is the binding-constraint principle, and it is the lens for every investment decision below.

Why Programs Fail

Beneath the failure statistics sit three recurring patterns, all of them leadership patterns:

  • Funding the visible layer, not the binding one. Money flows to the fashionable track — usually a platform or a model — while the actual constraint (often data, talent, or adoption) goes unaddressed. The spend looks like progress and moves nothing, because value is gated by the weakest dependency, not the largest line item. This is why "we bought the platform" and "we trained everyone" so rarely move the needle alone.
  • Tipping too sequential or too parallel. Too sequential ("waterfall transformation") finishes Strategy, then Governance, then Data, then Platform, and ships nothing to a real user for two years until executive patience runs out. Too parallel ("boil the ocean") stands up all eight tracks at once, thrashing on coordination until foundation gaps surface under work already built on top of them. The synthesis is to run the tracks concurrently but at staggered intensity, pouring resources into a track only once its upstream is ready.
  • Measuring nothing, so evidence can't steer. In McKinsey's survey, fewer than one in five organizations track well-defined KPIs for their gen-AI solutions (McKinsey, 2025). KPI tracking is also among the practices most correlated with bottom-line impact. Both halves are respondents' own accounts, collected by a firm that sells this work. Without instrumentation captured before deployment, value can't be proven, weak bets can't be killed early, and the program runs on anecdote until the budget is gone.

What Leadership Should Do First

The corrective is not faster spending; it is a disciplined start. The 90-day launch sequences it in full, but the leadership moves are four:

  1. Put one accountable executive over the whole program. In McKinsey's survey, CEO oversight of AI governance was one of the elements most correlated with self-reported bottom-line impact from gen AI, and the strongest at larger companies (McKinsey, 2025). Yet only 28% of respondents said their CEO oversees AI governance (McKinsey, 2025). Name a single owner who reports high enough to reallocate capital — and govern the transformation through a steering structure with explicit decision rights (see program architecture).
  2. Assess all eight tracks and find the binding constraint. Run the integrated assessment — a single 90-minute, cross-functional, evidence-based scoring of the whole board — and let it name the one constraint to relieve next, rather than funding by intuition or volume.
  3. Instrument before you deploy. Capture the pre-AI baseline and stand up measurement on day one, so the first win is provable and the first failure is cheap. The baseline you skip is the proof you will never have.
  4. Ship one measured win, then let evidence pull the next bet. Resist the portfolio of unmeasured pilots. Deliver one or two visible, attributable use cases tied to the constraint, measure them against the baseline, and reallocate on the result. Promise what you can measure — the credibility of the numbers is what keeps the program funded.

The throughline. Direct effort at the binding constraint, run the tracks at staggered intensity, and let measurement — not enthusiasm — pace the next investment. Spend follows evidence, not the loudest function. BCG's rule of thumb puts roughly 70% of the strategic focus on people and process, 20% on technology, and 10% on algorithms (BCG, 2025); budgets that invert that ratio are the gap, restated.

Key Takeaways

  • Adoption is near-universal; value is rare. In McKinsey's survey, 89% use AI but only ~6% attribute 5%+ of EBIT to it. Roughly 95% of pilots show no P&L impact (MIT NANDA, 2025), and only ~5% reach value at scale (BCG, 2025). Every figure here is self-reported to an organization with a stake in the answer. Companies BCG classes as leaders show far better results (BCG, 2025), but they are the exception.
  • Treat it as eight interdependent tracks, not one program. Direction → Foundation → Execution → Feedback; a track is only as strong as its weakest upstream dependency.
  • Most failure is a leadership pattern: funding the visible layer instead of the binding one, tipping too sequential or too parallel, and measuring nothing.
  • Start with discipline, not speed: one accountable executive, an integrated assessment to find the binding constraint, measurement instrumented before deployment, and one measured win that pulls the next bet.
  • Let evidence pace the spend. BCG's rule of thumb puts ~70% of the focus on people and process (BCG, 2025); fund the constraint, prove the value, and reallocate quarterly on what the measurement shows.

Sources

  • McKinsey — The state of AI in 2026: On the road to ROI, 2026. Nearly nine in ten respondents report regular use of AI in at least one business function. View source · verified 2026-09-13 · primary
  • McKinsey — The state of AI in 2026: On the road to ROI, 2026. About four in ten respondents (37 percent) report that AI has contributed positively to their organizations' EBIT, essentially unchanged from 2025. View source · verified 2026-09-13 · primary
  • McKinsey — The state of AI in 2026: On the road to ROI, 2026. AI high performers - respondents who attribute an EBIT impact of 5 percent or more to AI use and say their organizations have seen 'significant' value from AI use - account for just 6 percent of survey respondents, unchanged from 2025. View source · verified 2026-09-13 · primary
  • US Census Bureau — Large Firms With at Least 20 Employees Biggest AI Users (Business Trends and Outlook Survey), 2026. overall AI usage hovered between 17% and 20% ... the national rate (19.8%) ... 37% of firms with at least 250 employees reported using AI. View source · verified 2026-09-16 · primary
  • Eurostat — 20% of EU enterprises use AI technologies (survey on ICT usage and e-commerce in enterprises), 2025. 20.0% of EU enterprises with 10 or more employees used artificial intelligence ... 13.5% in 2024. View source · verified 2026-09-16 · primary
  • Bonney, Breaux, Dinlersoz, Foster, Haltiwanger & Pande (US Census Bureau CES) — The Microstructure of AI Diffusion: Evidence from Firms, Business Functions, and Worker Tasks (CES-WP-26-25), 2026. 18% of firms used AI in a business function, rising to 32% on an employment-weighted basis ... 50%-60% (60%-70%, employment-weighted) for very large firms ... 57% of users integrate AI in three or fewer business functions ... robust positive correlation between firm commercial performance and the breadth of AI integration. View source · verified 2026-09-16 · primary
  • Gartner — Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept By End of 2025, 2024. At least 30% of generative AI projects will be abandoned after proof of concept by the end of 2025, due to poor data quality, inadequate risk controls, escalating costs or unclear business value. View source · verified 2026-06-20 · primary
  • MIT Project NANDA — The GenAI Divide: State of AI in Business 2025, 2025. Just 5% of integrated AI pilots are extracting millions in value, while the vast majority remain stuck with no measurable P&L impact. View source · verified 2026-06-20 · ⚠ secondary mirror
  • BCG — The Widening AI Value Gap, 2025. Only 5% of companies in our 2025 study of more than 1,250 firms worldwide are achieving AI value at scale ... Fully 60% of companies are not achieving material value at all, reporting minimal revenue and cost gains despite substantial investment. Another 35% (13 percentage points more than in 2024) are scaling up their efforts and seeing some returns ... the 60% that have little or no value. View source · verified 2026-09-13 · primary
  • Cisco — 2025 Cisco AI Readiness Index: Realizing the Value of AI, 2025. The Most AI-ready Companies Outpace Peers in the Race to Value ... A small but consistent group of companies surveyed - the 'Pacesetters,' about 13% of organizations for the last three years - outperform their peers across every measure of AI value ... The 'Pacesetters' are 4x more likely to move AI pilots into production. View source · verified 2026-09-13 · primary
  • McKinsey — Superagency in the Workplace, 2025. Only 1 percent of leaders call their companies mature on the deployment spectrum, meaning AI is fully integrated into workflows and drives substantial business outcomes. View source · verified 2026-06-20 · primary
  • BCG — The Widening AI Value Gap, 2025. Future-built companies already generate 1.7 times more revenue growth and 1.6 times higher EBIT margins than the 60% of companies in the categories we term stagnating or emerging. ... As a result of this investment, they expect twice the revenue increase and 1.4 times greater cost reductions than laggards in the areas where they apply AI. View source · verified 2026-09-14 · primary
  • McKinsey — The State of AI: How Organizations Are Rewiring to Capture Value, 2025. Less than one in five organizations are tracking KPIs for gen AI solutions — and tracking well-defined KPIs is the practice with the most impact on the bottom line. View source · verified 2026-06-20 · primary
  • McKinsey — The State of AI: How Organizations Are Rewiring to Capture Value, 2025. Our survey analyses show that a CEO's oversight of AI governance ... is one element most correlated with higher self-reported bottom-line impact from an organization's gen AI use. ... That's particularly true at larger companies, where CEO oversight is the element with the most impact on EBIT attributable to gen AI. Twenty-eight percent of respondents whose organizations use AI report that their CEO is responsible for overseeing AI governance ... and 17 percent say AI governance is overseen by their board of directors. View source · verified 2026-09-14 · primary
  • BCG — The Widening AI Value Gap, 2025. Adherence to our 10-20-70 rule for technology transformations will help speed the journey: 70% of a business's strategic focus should be on the people and processes, 20% on the tech, and 10% on algorithms. View source · verified 2026-09-13 · primary