A cross-industry analysis of where enterprise AI spending goes, where value is actually created, and what the 6% generating EBIT do differently.
Enterprise AI spending is projected to reach $2.52 trillion in 2026, a 44% increase over 2025. Yet across six independent research programs—McKinsey, BCG, Deloitte, Gartner, MIT, and Prosci—a consistent pattern emerges: the vast majority of organizations are not generating measurable returns on this investment. Only 6% of organizations report that AI contributes 5% or more to EBIT. Ninety-five percent of enterprise AI pilots deliver no measurable P&L impact.
The evidence points to a structural explanation. Organizations are allocating 93% of AI budgets to technology and 7% to the people, processes, and organizational design required to absorb it. This ratio is inverted relative to what the evidence says drives returns. BCG estimates that AI success depends approximately 70% on people and processes, 20% on technology infrastructure, and 10% on algorithms.
Four independent research streams—budget allocation data, implementation failure analysis, organizational absorption studies, and change management ROI research—converge on a single insight none of them state individually: the organizations generating returns from AI are not spending more on technology. They are spending differently. The technology works. The organization cannot absorb it.
This brief extends the analysis in 3 Foundational Decisions That Separate the 6% Generating EBIT from AI (SI-Q4), which identified the governance, ownership, and portfolio decisions that distinguish the top performers. Where that brief asks what to decide, this one asks how to fund it—examining the budget architecture, organizational absorption capacity, and investment sequencing that turn those decisions into measurable returns.
Enterprise AI spending is accelerating at a pace without precedent in corporate technology investment. Gartner forecasts worldwide AI spending will total $2.52 trillion in 2026, a 44% increase from 2025. IDC projects enterprise AI solutions spending alone will reach $307 billion in 2025, growing to $632 billion by 2028. The infrastructure layer is even larger: AI infrastructure spending reached $82 billion in Q2 2025 alone, a 166% year-over-year increase.
This capital is flowing from every direction. One-third of companies surveyed by BCG plan to commit more than $25 million to AI initiatives in 2025. Seventy-five percent of CFOs plan to increase technology budgets in 2026, with nearly half planning increases of 10% or more. The investment thesis is no longer in question. Eighty-eight percent of organizations now use AI in at least one business function, up from 78% a year ago.
The scale of commitment is not the question. The question is whether these investments are structured to generate returns.
The data on AI returns tells a remarkably consistent story across every major research program that has measured it. McKinsey's 2025 State of AI survey, covering 1,933 participants, found that only about 6% of respondents attribute 5% or more of their organization's EBIT to AI. More than 80% report no tangible impact on enterprise-level EBIT from generative AI.
BCG's AI Radar 2025, surveying 1,803 C-level executives across 19 markets, found that only a quarter report meaningful value from their AI initiatives. Meanwhile, a Gartner survey of 506 CIOs found that 72% of organizations are breaking even or losing money on AI investments.
MIT's NANDA initiative published the most granular analysis to date. Their "GenAI Divide" report, based on 52 executive interviews, surveys of 153 leaders, and analysis of 300 public AI deployments, found that 95% of enterprise AI pilots delivered no measurable P&L impact. Only 5% of integrated systems created significant value.
| Source | Sample | Finding | Year |
|---|---|---|---|
| McKinsey State of AI | 1,933 respondents | Only 6% attribute 5%+ EBIT to AI | 2025 |
| BCG AI Radar | 1,803 C-level executives | Only 25% report meaningful AI value | 2025 |
| Gartner CIO Survey | 506 CIOs | 72% breaking even or losing money | 2025 |
| MIT NANDA | 300 deployments analyzed | 95% of pilots: no P&L impact | 2025 |
The consistency across these independent studies is striking. They use different methodologies, different sample populations, and different definitions of success. Yet all four converge on the same structural conclusion: the gap between AI investment and AI returns is not narrowing. It is widening.
The pattern holds across industries, though the mechanisms differ. In financial services, regulatory and cybersecurity requirements absorb a disproportionate share of AI budgets—74% of financial services respondents in Deloitte's survey plan to increase cybersecurity spending alongside AI, compressing the already thin allocation for organizational capability. In manufacturing and CPG, the constraint is different: legacy ERP and OT environments require extensive integration work before AI can reach production workflows, locking capital into technology interoperability rather than process redesign. In professional services, where the asset is intellectual capital rather than physical infrastructure, the challenge is demonstrating AI ROI without a traditional cost-of-goods-sold structure—partnership economics make it structurally difficult to fund organizational transformation when the budget case must be made partner by partner.
Sixty percent of companies lack defined financial KPIs for their AI initiatives (BCG AI Radar, 2025). Organizations are spending without a measurement framework—the equivalent of running a clinical trial without defining what success looks like before starting.
In December 2025, Deloitte's CTO Bill Briggs disclosed a statistic that crystallizes the structural problem. Organizations are allocating 93% of their AI budgets to technology—models, chips, infrastructure, software—and 7% to the people expected to use it. Briggs called this a critical error, likening it to obsessing over "the ingredients" while ignoring "the recipe."
This 93/7 ratio stands in direct opposition to what the research says drives returns. BCG's AI Radar 2025 estimates that AI success depends approximately 70% on people and processes, 20% on technology infrastructure, and only 10% on algorithms and models. The gap between how organizations spend and where value is created is not a rounding error. It is a 13:1 ratio inversion.
Deloitte's own enterprise survey reinforces the point from the demand side: 84% of companies have not redesigned jobs around AI capabilities. Despite increasing access to generative AI in the workplace, overall usage has actually decreased by 15%. The technology is available. The organization has not been prepared to use it.
This is not a communication problem or a training problem. It is a budget architecture problem. When 93 cents of every AI dollar go to technology acquisition and 7 cents go to everything else—workflow redesign, role reconstruction, governance design, change management—the result is predictable. The technology arrives. The organization cannot absorb it.
If the technology works and the returns are missing, what are the organizations generating EBIT doing differently with the same technology?
No single research program identifies what separates the 6% from the 94%. But when four independent evidence streams are placed side by side, a convergence pattern emerges that none of them articulate individually.
Deloitte reports the 93/7 technology-to-people spending ratio. Gartner notes that CFOs are increasing technology budgets by 10% while cutting HR budgets—only 29% plan to increase HR spending in 2026, and 22% expect cuts. The budget architecture structurally underweights the capability that drives adoption.
MIT's NANDA study found that 95% of AI pilots deliver no P&L impact. But the study also found that purchasing AI tools from specialized vendors succeeds 67% of the time, while internal builds succeed only one-third as often. The technology itself is not the failure point. The organizational integration is.
McKinsey's 6% high performers are 3.6 times more likely to be aiming for transformational, enterprise-level change rather than incremental improvements. Fifty-five percent of them fundamentally reworked processes when deploying AI—nearly three times the rate of other organizations. BCG found that leading firms focus on 3.5 AI use cases versus 6.1 for lagging firms, yet generate 2.1 times more ROI. High performers concentrate investment and redesign work around AI. Others spread technology across unchanged processes.
Prosci's benchmarking data, drawing from over 2,600 change practitioners, shows that projects with excellent change management are 7 times more likely to achieve their objectives. The success rate is 88% with excellent change management versus 13% when change management is poor or absent. When the people side of change is not managed, the probability of meeting project objectives drops to 16%.
The convergence insight: The 6% of organizations generating EBIT from AI are not spending more on technology. They are inverting the ratio. They concentrate AI investment on fewer use cases, fundamentally redesign work processes around AI capabilities, and fund organizational change at a level commensurate with the transformation they are attempting. The technology is a constant. The variable is organizational absorption capacity—and absorption capacity is a budget line item, not a cultural aspiration.
Synthesizing the budget allocation data with the returns evidence, five distinct investment archetypes emerge across the organizations studied. Each reflects a different structural relationship between technology spending and organizational capability investment.
The following diagnostic helps leadership teams assess whether their AI investment structure is aligned with the evidence on what drives returns. It is organized around the Four Capability Bands from The Intelligence Organization™: Right-Fit Technology, People & Purpose, Operational Integration, and Adaptive Governance.
| Capability Band | Assessment Question | Score 1–5 | Budget Implication |
|---|---|---|---|
| Band 1: Right-Fit Technology | Are AI investments concentrated on 3–4 use cases with clear value thresholds, or spread across 6+? | ___ | Score <3: freeze new acquisitions until the current portfolio is concentrated. Scattered investments are the single most common pattern in the 95% of pilots that deliver no measurable P&L impact. |
| For each AI initiative, has the organization assessed whether the team has the skills, data infrastructure, and process maturity to actually operate it—or was the investment approved based on what the technology can do? | ___ | Score <3: the technology plan exceeds the organization’s capacity to absorb it. Score each investment against what the team can sustain, not what the vendor demonstrated. | |
| Band 2: People & Purpose | What percentage of AI budget funds workflow redesign, role reconstruction, and capability building versus technology licensing and infrastructure? | ___ | Score <3: reallocate minimum 20% of AI budget to people and process. The 93/7 ratio inverts where value is created—BCG estimates 70% of AI success depends on people and processes. |
| Has the organization identified who is already using AI informally and willing to lead adoption, versus who needs demonstrated proof before engaging—and are those groups being supported differently? | ___ | Score <3: a uniform adoption strategy (training everyone the same way) wastes the budget. Early adopters need autonomy and fast access; skeptics need evidence from peers, not mandates from leadership. | |
| Band 3: Operational Integration | Have work processes been fundamentally redesigned around AI, or has AI been layered onto existing workflows without changing how work is structured? | ___ | Score <3: identify where people waste time on manual translation, cross-system reconciliation, or redundant data entry—those are the highest-value redesign targets, not the most visible ones. |
| When one process is redesigned, are the upstream and downstream processes that depend on it redesigned together—or does optimizing one workflow just move the bottleneck to the next? | ___ | Score <3: redesign connected processes as a group. Automating one step in isolation speeds up that step but creates friction at every handoff. | |
| Band 4: Adaptive Governance | Do AI financial KPIs exist, are they measured, and does governance adapt based on results? | ___ | Score <3: define financial KPIs before approving new AI spend. Sixty percent of organizations cannot measure AI ROI—governance without measurement is governance without accountability. |
| Does every AI initiative move through the same approval process regardless of risk level, or are low-risk internal tools pre-authorized to move fast while high-risk customer-facing applications get deliberate review? | ___ | Score <3: one governance speed for everything guarantees that either low-risk work moves too slowly or high-risk work moves too fast. Separate the lanes. |
Total 32–40: Capability Builder or Greenfield Designer archetype. Investment ratio likely already approaching 50/50 or better. Focus on sustaining and scaling.
Total 20–31: Process Redesigner archetype. Budget inversion has begun but is incomplete. Identify the lowest-scoring Band and address it before adding technology.
Total 8–19: Technology Collector or Pilot Factory archetype. Investment ratio likely at or near 93/7. Additional technology spending will not improve returns. Begin with Band 2 (People & Purpose)—the binding constraint for 84% of organizations.
The diagnostic applies differently depending on organizational starting position:
Post-merger integration: When budgets are locked into system consolidation, the diagnostic reveals that AI investment is structurally trapped in Band 1 (Right-Fit Technology). The priority is not adding AI spend but carving out a protected allocation for Bands 2–4 before integration budgets are fully consumed. Start with Band 3 (Operational Integration)—the integration itself is the process redesign opportunity.
Turnaround or cost pressure: Organizations under financial stress have a paradoxical advantage: the urgency to restructure creates executive permission to reallocate budgets that stable organizations resist. The diagnostic should target the 2–3 use cases with the shortest path to measurable cost reduction. Band 4 (Adaptive Governance) becomes the priority—financial KPIs must be defined before any AI dollar moves.
Greenfield or new entity: Organizations building from scratch should skip Bands 1 and 3 entirely. The technology layer is commoditized and the processes do not yet exist to redesign. All diagnostic energy goes to Band 2 (People & Purpose) and Band 4 (Adaptive Governance)—designing organizational capability before selecting technology, inverting the sequence that traps legacy firms.
Zero dedicated AI budget: When the organization has no formal AI allocation, the pilot trap (Archetype 02) is the default. The diagnostic reframes the business case: instead of requesting budget for AI technology, request budget for organizational capability with AI as the enabling mechanism. Band 2 scoring provides the evidence base for the investment request.
Decision support aligned with The Intelligence Organization · Band 2 (People & Purpose) as binding constraint · Absorption capacity as the primary determinant of AI ROIThis research is the foundation for our AI investment architecture executive workshop series. If your organization is evaluating how to restructure AI spending for returns, we should talk.
Schedule a ConversationIndustry Research: McKinsey & Company, "The State of AI in 2025: Agents, Innovation, and Transformation," Mar 2025, 1,933 respondents. BCG, "From Potential to Profit: Closing the AI Impact Gap," BCG AI Radar, Jan 2025, 1,803 C-level executives across 19 markets. BCG, "From Potential to Profit with GenAI," BCG AI Radar, Jan 2024, 1,406 executives in 50 markets. Deloitte, "The State of AI in the Enterprise, 2026," 3,235 business and IT leaders across 24 countries. Deloitte, Bill Briggs (CTO), Fortune interview, Dec 15 2025. Gartner, "Worldwide AI Spending Will Total $2.5 Trillion in 2026," Jan 15 2026. Gartner, "CFOs' Budget Plans Prioritize Growth Functions, Technology and AI in 2026," Feb 10 2026. Gartner CIO Survey, May 2025, 506 CIOs. Gartner, Alicia Mullery, "All IT Work Will Involve AI by 2030," IT Symposium, Oct 2025.
Academic & Institutional: MIT NANDA Initiative, "GenAI Divide: State of AI in Business 2025," Aug 2025, 52 executive interviews, 153 leader surveys, 300 deployment analyses. IDC, "Worldwide Spending on Artificial Intelligence Forecast to Reach $632 Billion in 2028," 2024. IDC, "Artificial Intelligence Infrastructure Spending to Reach $758 Billion by 2029," Q2 2025.
Change Management & Organizational: Prosci, "Best Practices in Change Management," 11th & 12th Editions, 2,600+ change practitioners. Prosci, "The Correlation Between Change Management and Project Success."
RBD. Research: Starkey, M.C., The Intelligence Organization, 2026. RBD., "3 Foundational Decisions: What Separates the 6% Generating EBIT from AI," SI-Q4 2025. RBD., "Enterprise AI Tool Strategy: The Copilot Adoption Gap," SI-Q1 2026. RBD. cross-industry AI investment and returns synthesis, 2026.