The Maturity Cliff: Why 62% of Companies Are Below the Financial Performance Threshold

See also (wiki): ai-maturity-models · workflow-redesign · ai-roadmap-execution

Executive Summary

• MIT CISR’s four-stage Enterprise AI Maturity Model, built from 721 companies (2022) and updated with 152 companies and 19 executive interviews (2025), identifies a hard financial break between stages 2 and 3: organizations in stages 1–2 run below industry-average financial performance; those in stages 3–4 run well above it.
• The specific gap: Stage 3 organizations post +11.3 pp growth and +8.7 pp profit above their industry average. Stage 1 organizations post −12.6 pp growth and −9.6 pp profit below theirs. That is a 23.9-point growth spread and an 18.3-point profit spread between the top and bottom stages.
• As of 2022, 62% of enterprises sat in stages 1 or 2 — building pilots, running experiments, but not yet scaling. By 2025, 64% had moved to stages 3 or 4. Progress is real, but most companies are still early in capturing the financial dividend.
• The transition from stage 2 to stage 3 is the highest-value move a company can make. Stage 2 organizations have demonstrated AI value. Stage 3 organizations have industrialized it — which requires workflow architecture, not more pilots.
• The stage 4 minority (18% by 2025) is doing something qualitatively different: embedding AI in all decisions and building AI-augmented services to sell to others. That is a competitive moat, not just an efficiency play.

The Model: Four Stages, One Financial Break

MIT CISR is the Center for Information Systems Research at MIT Sloan School of Management. The Enterprise AI Maturity Model was first published in December 2024 (Weill, Woerner, Sebastian) drawing on a 2022 survey of 721 companies. An August 2025 update by Woerner, Sebastian, Weill, and Káganer tracked progress against the same framework using 152 companies and 19 executive interviews across 11 enterprises.

Source credibility: HIGH. MIT CISR is a non-vendor academic research center. The 2022 base survey (n=721) is large and cross-industry. The 2025 update (n=152) is smaller but supplemented by substantive executive interviews. Financial performance is measured as deviation from industry average — a relative benchmark that controls for sector differences. Limitation: the 2022 survey is three years old in a fast-moving domain; the 2025 update is more current but smaller. The model structure is directionally robust; exact percentages should be treated as indicative.

Stage 1: Experiment and Prepare (28% in 2022 → 13% in 2025)

Organizations educate leadership, establish acceptable use policies, and run initial experiments. AI literacy exists at the executive level; adoption is not yet systematic. Financial result: −12.6 pp growth and −9.6 pp profit versus industry average. These companies are spending on AI without yet realizing financial returns.

Stage 2: Build Pilots and Capabilities (34% in 2022 → 23% in 2025)

Pilots are running and demonstrating value. Processes are being simplified. Data is being consolidated for AI consumption. LLMs are in use. But adoption is not scaled — it depends on individual teams and project sponsors, not enterprise-wide systems. Financial result: −3.5 pp growth and −2.2 pp profit versus industry average. Improvement from Stage 1, but still below average. The insight here is important: proven pilots do not produce above-average financial performance. Scaling does.

Stage 3: Develop AI Ways of Working (31% in 2022 → 46% in 2025)

AI is industrialized across the enterprise. Reusable architecture exists. A test-and-learn culture is operating at scale, not as an experiment. Foundation and small language models are applied to proprietary data. Dashboards track outcomes, not just adoption. Financial result: +11.3 pp growth and +8.7 pp profit versus industry average. This is where the financial break occurs.

The 2025 CISR briefing calls the Stage 2→3 transition “the greatest financial impact” move. What makes it hard is that it is not a technology purchase. It requires four organizational shifts that CISR labels Strategy, Systems, Synchronization, and Stewardship: aligning AI investments to measurable strategic goals; building modular, interoperable data platforms; redesigning roles and teams for AI; and embedding compliance and transparency by design.

Stage 4: Become AI Future-Ready (7% in 2022 → 18% in 2025)

AI is embedded in all decision-making. The organization has developed proprietary AI capabilities and, critically, is building AI-augmented services to sell externally. This is where AI becomes a revenue source, not just a cost reduction lever. Financial result: +17.1 pp growth and +10.4 pp profit versus industry average.

The Stage 3→4 gap is narrower than the Stage 2→3 gap. Moving from Stage 3 to Stage 4 adds roughly 5.8 pp of growth and 1.7 pp of profit. The biggest single move is getting to Stage 3.

The Performance Gap in Numbers

Total spread (Stage 1 to Stage 4): 29.7 pp of growth, 20 pp of profit.

The migration upward between 2022 and 2025 is notable: Stage 1 shrank by 15 points, Stage 4 grew by 11 points. But 36% of companies remain in stages 1–2 as of 2025. For those companies, the financial cost of staying there is measurable and growing.

What Stage 3 Actually Requires

The CISR research is specific about why Stage 2 companies stall. The bottleneck is not technology. It is organizational architecture.

Reusable infrastructure, not one-off builds. Stage 3 requires modular, interoperable platforms so that AI built for one process can be deployed across others. Stage 2 organizations build pilots that work in isolation. Stage 3 organizations build systems that scale.

Outcomes dashboards, not adoption metrics. Stage 2 organizations count users and use cases. Stage 3 organizations measure business results — cycle time, defect rate, revenue per customer — and tie AI directly to those numbers. Adoption metrics confirm deployment; outcome metrics confirm value.

Role redesign, not tool distribution. The CISR framing on workforce (“synchronization”) is explicit: Stage 3 requires redesigned roles and teams, not just employees with new tools. The hardest organizational change is moving from command-and-control management to coaching-and-outcome management — because AI-assisted employees need autonomy to use it, and traditional supervisory structures often block that.

Stewardship built in, not layered on. Compliance and transparency requirements must be embedded in the architecture from Stage 3 forward. Organizations that treat governance as an audit exercise rather than a design constraint spend the Stage 3 transition fighting compliance reviews rather than scaling.

Case study data from CISR supports the pattern. Guardian Life Insurance automated its RFP and quoting process, cutting turnaround from one week to 24 hours — a Stage 3 scaling result. Italgas accelerated construction workflows 40%, reduced physical inspections 80%, and generated €3 million in 2024 revenue by commercializing the resulting capabilities — a Stage 4 play.

Key Data Points

What This Means for Your Organization

The CISR model is useful precisely because it separates activity from outcome. Many organizations have activity: executives trained, policies written, pilots running, some user adoption. CISR’s data shows that activity at Stages 1 and 2 correlates with below-average financial performance — not because AI is not working, but because the organizational work of scaling has not yet happened.

The question this research poses for any company currently in active AI pilots is: what is your plan to get to Stage 3? Not “launch another pilot” — but to build the reusable infrastructure, redesign the workflows, and change the management model that makes the existing pilots repeatable across the enterprise. The Stage 2→3 transition is an organizational project, not a technology project. Most technology vendors will not tell you that.

The 2025 migration data is worth absorbing: in three years, the percentage of companies at Stage 3 or above grew from 38% to 64%. The companies that moved are already capturing the financial premium. The companies that stayed at Stage 2 are falling further behind on a relative basis — not because they did nothing, but because their competitors moved.

For organizations trying to honestly assess where they sit in this model and what the Stage 3 transition actually requires in their specific context, that conversation is more useful than most benchmark exercises: brandon@brandonsneider.com.

Sources

1. Weill, P., Woerner, S.L., & Sebastian, I.M. (2024, December). “Building Enterprise AI Maturity.” MIT CISR Research Briefing No. XXIV-12. n=721 companies (October 2022 survey) + 16 executive interviews across 9 enterprises. Credibility: HIGH. MIT non-vendor academic research center; large cross-industry sample; financial performance measured relative to industry average controls for sector effects. Limitation: 2022 data is three years old in a fast-moving domain.

   • URL: https://cisr.mit.edu/publication/2024_1201_EnterpriseAIMaturityModel_WeillWoernerSebastian

2. Woerner, S.L., Sebastian, I.M., Weill, P., & Káganer, E. (2025, August 21). “Grow Enterprise AI Maturity for Bottom-Line Impact.” MIT CISR Research Briefing No. XXV-8. n=152 companies (2025 survey) + 19 executive interviews across 11 enterprises (2024–2025). Credibility: HIGH. Updates the 2024 model with current distribution data; smaller sample but substantive executive interviews. Key finding: greatest financial impact occurs in Stage 2→3 transition.

   • URL: https://cisr.mit.edu/publication/2025_0801_EnterpriseAIMaturityUpdate_WoernerSebastianWeillKaganer

3. MIT Sloan School of Management. (2024, December 19). Press release: “New MIT CISR research finds companies with advanced enterprise AI outpace industry peers in financial performance.” Used for cross-verification of key statistics and quote attribution.

   • URL: https://mitsloan.mit.edu/press/new-mit-cisr-research-finds-companies-advanced-enterprise-ai-outpace-industry-peers-financial-performance

4. MIT Sloan Management Review. (2025). “What’s your company’s AI maturity level?” Summary for business audience; confirms stage descriptions and financial performance findings. Credibility: MEDIUM-HIGH (secondary summary of primary research).

   • URL: https://mitsloan.mit.edu/ideas-made-to-matter/whats-your-companys-ai-maturity-level

5. MIT Sloan Management Review. (2025). “How to boost your organization’s AI maturity level.” Covers four organizational challenges for Stage 3 scaling and Guardian Life/Italgas case studies. Used for organizational capability detail. Credibility: MEDIUM-HIGH (secondary summary with case study data).

   • URL: https://mitsloan.mit.edu/ideas-made-to-matter/how-to-boost-your-organizations-ai-maturity-level

Brandon Sneider | brandon@brandonsneider.com March 2026