Five consulting firms. Five independent research programmes. Five separate datasets, methodologies, and sample populations. And the same conclusion: enterprise AI investment has reached unprecedented levels, but the return on that investment remains concentrated in a small minority of organisations. The rest are spending more, deploying more, reporting more — and getting less than they expected.
This article synthesises the core findings from McKinsey's State of AI 2025, BCG's AI Radar 2025, Accenture's Technology Vision 2025, Deloitte's State of Generative AI in the Enterprise, and Bain's Generative AI survey. Each study approaches the question from a different angle. Together, they form the most comprehensive empirical picture available of why the investment-returns gap exists and what closes it.
The scale of the gap
McKinsey quantifies the ceiling and the floor. The McKinsey Global Institute's estimate of $2.6 to $4.4 trillion in annual value potential across 63 use cases remains the benchmark for what AI could deliver. The 2025 State of AI survey shows what it actually delivers: 88% of organisations use AI, but only 6% achieve more than 5% EBIT impact. That 6% figure represents 109 companies out of 1,993 surveyed. The remaining 94% — including the majority that have made substantial AI investments — see returns that do not register on the income statement in a meaningful way. The full McKinsey analysis examines this gap in detail.
BCG draws the same line with different data. BCG's "Build for the Future" study surveyed 1,250 senior executives across 9 industries and more than 25 sectors, assessing AI maturity across 41 foundational capabilities. Only 5% qualify as "future-built" — BCG's term for organisations that have structurally integrated AI into their operating model. The performance differential is stark: future-built firms achieve 1.7 times revenue growth, 3.6 times total shareholder return, and 1.6 times EBIT margin improvement compared to laggards. These are not incremental advantages. They represent a different trajectory. Future-built firms also invest more than twice as much in AI as their peers — but the investment follows the structural transformation rather than preceding it. The BCG maturity blueprint breaks down the capability architecture that separates these tiers.
Accenture measures the acceleration. Accenture's Reinventors — companies that have fundamentally reorganised operations around technology — outperform their peers by 15 percentage points on multiple financial metrics, a gap Accenture projects will widen to 37 percentage points by 2026. With $36 billion in generative AI bookings, Accenture's own commercial data confirms the scale of enterprise spending. The Reinventors' advantage is not that they spend more. It is that their spending compounds because it builds on structural changes rather than adding tools to unchanged processes. The reinvention premium analysis explores what distinguishes Reinventors from the rest.
Deloitte captures the paradox of progress. Deloitte's survey reveals a pattern that contradicts the optimistic adoption narrative: 25% of organisations have moved 40% or more of their AI experiments into production, and the share reporting transformative impact has doubled from 12% to 25% year over year. Progress is real. But here is the paradox — perceived preparedness is declining simultaneously. Only 43% rate their technical infrastructure as ready. Data management readiness sits at 40%. Talent readiness has dropped to 20%. Organisations that have deployed AI at scale are not becoming more confident. They are discovering that scale introduces complexities that were invisible during the pilot phase. More AI in production does not automatically mean more AI creating value.
Bain documents the breadth without depth. Bain's survey of US enterprises reports that 95% now use generative AI, and the number of use cases per company has doubled in fourteen months. The expansion is broad but shallow. Bain finds that meaningful EBITDA gains — in the range of 10 to 25% — occur only in technology-forward enterprises that have fundamentally redesigned workflows around AI capabilities. The majority of companies using generative AI across many functions see productivity improvements that are real at the individual level but do not translate into financial outcomes at the enterprise level.
The five-study convergence
The common denominator is not technology spending. If it were, the gap would be closing as investment increases. It is not. The common denominator is structural transformation — and all five studies describe it in remarkably similar terms despite their different frameworks.
Workflow redesign, not tool deployment. McKinsey finds that high performers are 2.8 times more likely to have redesigned workflows around AI rather than layering AI onto existing processes. BCG's future-built firms have restructured their operating models. Accenture's Reinventors have reorganised operations. Bain specifies workflow redesign as the prerequisite for EBITDA impact. Deloitte's data shows that production deployment without workflow change produces scale without proportional returns. The vocabulary differs. The finding is identical.
Business ownership, not IT ownership. McKinsey reports that high performers' leaders are three times more likely to demonstrate active ownership of AI initiatives. BCG's future-built category requires C-suite integration of AI into strategic planning, not delegation to technology functions. Accenture's Reinventors are led by executives who treat technology transformation as a business strategy, not a technology project. Bain's technology-forward enterprises assign P&L-accountable leaders to AI programmes. When AI reports to IT, it optimises technology. When it reports to business leadership, it optimises outcomes.
Governance and risk management, not governance theatre. McKinsey finds that high performers invest in monitoring and governance infrastructure while the 94% rely on policies. BCG's maturity model includes operational governance as a defining characteristic of future-built firms. Deloitte documents declining preparedness despite increasing deployment, suggesting that governance has not scaled with adoption. Accenture defines trust — built on accuracy, predictability, consistency, and traceability — as the precondition for realising AI's full benefits, with 77% of executives affirming this position. The trust barrier analysis examines why governance is the operational bottleneck, not a compliance formality.
Capability building, not tool purchasing. Every study distinguishes between organisations that buy AI tools and organisations that build AI capabilities. The distinction is not semantic. Tools are software products that individual employees use. Capabilities are organisational muscles — data infrastructure that makes enterprise knowledge accessible to AI systems, evaluation frameworks that measure AI performance against business outcomes, feedback loops that improve AI systems based on production data, and talent that can design, deploy, and manage AI workflows rather than simply use AI interfaces.
Why the gap is widening
The compounding effect explains the divergence. BCG's 3.6 times total shareholder return differential is not a one-year advantage. It is the result of compounding: each AI-enabled workflow improvement creates data that improves the next workflow, builds organisational muscle that accelerates the next deployment, and generates returns that fund the next investment. The 5-6% that have crossed the structural threshold are accelerating. The rest are iterating on pilots.
The talent constraint amplifies the divide. Deloitte's finding that talent readiness has dropped to 20% reflects a real structural problem. The organisations that transformed early attracted and developed the talent that enables continued transformation. The organisations that delayed face a market where AI engineering, data architecture, and AI product management skills are scarce and expensive. Talent follows impact — and impact is concentrating in the organisations that already have it.
The governance burden scales non-linearly. As organisations deploy more AI systems, the governance requirements multiply. Each system needs monitoring, each workflow needs delegation rules, each model update needs validation. Organisations that built governance infrastructure early — as part of their first deployments — have a scalable foundation. Organisations that deployed first and are now retrofitting governance face exponentially growing complexity. The cost structure analysis shows how governance costs compound when treated as an afterthought.
What the synthesis means for DACH enterprises
The investment-returns gap is not an American phenomenon observed from a distance. DACH enterprises face the same dynamics with two additional constraints: the EU AI Act imposes governance requirements that other markets do not face, and the Mittelstand's traditional strength — operational excellence within established processes — becomes a liability when those processes must be redesigned rather than optimised.
The path from the 94% to the 6% is not more AI. It is a different operating model for AI. The five studies converge on the same structural requirements: workflows redesigned around AI capabilities rather than augmented with AI tools, business leaders who own AI outcomes rather than technology leaders who own AI infrastructure, governance architecture that enables scaling rather than policy documents that constrain it, and capability building that creates organisational muscle rather than tool procurement that creates software licences.
The question is not whether to invest in AI. That decision has already been made — 88 to 95% of enterprises are investing. The question is whether the investment is structural or incremental. Structural investment changes how work flows, how decisions are made, and how the organisation learns from AI-generated data. Incremental investment adds AI to existing processes and hopes the productivity gains aggregate into financial impact. Five independent studies, covering tens of thousands of organisations, confirm that they do not.
Every article in this series has examined one dimension of the investment-returns gap — the McKinsey scaling analysis, the BCG maturity framework, the Accenture reinvention premium, the trust infrastructure required to scale, and the cost structures that determine whether investment compounds or dissipates. This synthesis makes the unified argument: the gap is real, it is widening, and closing it requires not more AI investment but a fundamentally different approach to how that investment is deployed.
Book a Fit Call to evaluate whether your AI approach is structural or incremental. We assess your current investment pattern against the five-study framework and identify the specific structural gaps — workflow design, ownership model, governance architecture, or capability foundation — that determine whether your AI spend will compound or plateau. Book your Fit Call →
References: McKinsey & Company, "The State of AI: How Organizations Are Rewiring to Capture Value," Global Survey, November 2025; BCG, "Build for the Future: Widening AI Value Gap," September 2025; Accenture, "Technology Vision 2025" and "Reinvention in the Age of Generative AI"; Deloitte, "State of AI in the Enterprise," 2026 edition; Bain & Company, "Generative AI and the Enterprise: From Pilots to Production," US Enterprise Survey, 2025.