← Knowledge Base 🕐 52 min read

Knowledge Base

AI Maturity Models

· March 2026

The structured frameworks — usually 3–5 stages — that measure where an organization sits along the arc from first AI ...

The structured frameworks — usually 3–5 stages — that measure where an organization sits along the arc from first AI experiment to AI-embedded business model. Maturity stage correlates strongly with financial performance: MIT CISR’s four-stage model (n=721, 2022; n=152 update, 2025) shows a hard financial break between stages 2 and 3. Stage 1 organizations run 12.6 points below industry-average growth and 9.6 points below industry-average profit; Stage 3 organizations run 11.3 points above growth and 8.7 points above profit. That is a 23.9-point growth spread and 18.3-point profit spread between bottom and top stages.

Why this matters to mid-market buyers

The Stage 2→3 transition is the highest-value move a company can make. Stage 2 organizations have demonstrated AI value in pilots. Stage 3 organizations have industrialized it. The move requires workflow architecture, not more pilots — which is why 62% of enterprises (2022) and 36% (2025) remain stuck below the financial threshold.
Different frameworks agree on the shape of the curve. BCG (5% capture substantial financial gains), McKinsey (6% high performers with >5% EBIT impact), Deloitte (transform vs. process vs. surface tier), Accenture (8% front-runners), and MIT CISR (18% in Stage 4 by 2025) converge on a single pattern: a small minority captures most of the financial upside, a larger majority runs pilots without structural change, and the rest experiment without scaling.
Maturity is not a technology purchase. The Stage 2→3 gap is bridged by four organizational shifts MIT CISR labels Strategy, Systems, Synchronization, and Stewardship — aligning AI to measurable goals, building modular data platforms, redesigning roles, and embedding compliance by design. None of those are vendor line items.

IBM IBV “From AI Projects to Profits” (n=2,500, Jun 2025) — ROI by Maturity Stage

IBM’s n=2,500 Oxford Economics survey provides one of the few sources that tracks ROI evolution across the maturity arc in the same dataset:

Ad hoc / peripheral: early returns ~31% (2023); current average ~7% — below cost of capital
Core-function deployment (top decile): ~18% — above cost of capital
The performance differential is not from the model but from deployment architecture: the top decile has horizontal workflows, interoperable data, and AI governance in place.
6% still ad hoc (down from 19%) — maturity is advancing, but fewer than a quarter have made the structural leap to AI-redesigned workflows.
Corroborates BCG’s 5%-capture-substantial-gains, McKinsey’s 6%-high-performers, and MIT CISR’s Stage 3 threshold findings from independent samples.

Source: research/12-agent-workers/ibm-ibv-ai-projects-to-profits-2025.md — MEDIUM-HIGH / TIER 2

Federal Reserve Adoption Baseline (Why Survey Numbers Conflict)

Every maturity-model discussion runs into the survey-figure conflict: McKinsey says 88% adoption; Fed says 18%. The Federal Reserve’s April 2026 FEDS Notes (Jeffrey S. Allen) resolves this with three-survey triangulation. The 18%–78% range reflects sampling, not disagreement: BTOS counts firms (95% small businesses); SBU weights by employment (large employers dominate). The most actionable single number for mid-market benchmarking is 41% — the RPS individual worker GenAI usage rate at work, nationally representative, growing at 31% per year.

Source: research/01-ai-native-landscape/fed-feds-notes-monitoring-ai-adoption-2026.md

Supporting research

research/01-ai-native-landscape/mit-cisr-enterprise-ai-maturity-2025.md — four-stage model with financial performance per stage; 721-company base survey (2022) + 152-company update (2025)
research/01-ai-native-landscape/mit-cisr-scaling-ai-maturity-bottom-line-2026.md — March 2026 MIT CISR session update with refreshed stage-progression and bottom-line data
research/01-ai-native-landscape/unified-ai-maturity-framework-synthesis.md — cross-maps MIT CISR (4 stages), BCG (3 tiers), McKinsey (high performer vs. POC), Deloitte (surface/process/transform), Stanford (automation intensity) into a single 3-stage model with financial outcomes at each stage
research/01-ai-native-landscape/bcg-ai-radar-2026.md — Trailblazer/Pragmatist/Follower archetypes (15%/70%/15%) with 2.7x–3.6x intensity gap across transformation budget, workforce upskilling, and CEO time (BCG, n=640 CEOs, Jan 15, 2026)
research/01-ai-native-landscape/mckinsey-state-of-ai-november-2025.md — 6% high performers capturing >5% EBIT impact; 1/3 scaling enterprise-wide, 2/3 still in POC
research/01-ai-native-landscape/mckinsey-state-of-ai-march-2025.md — March 2025 edition (Jul 2024 survey, n=1,491): workflow redesign is the single highest-EBIT-impact attribute of 25 tested; only 21% have redesigned any workflows; only 1% describe rollouts as “mature” — TIER 3 (superseded by Nov 2025 edition for current benchmarks, but the 25-attribute EBIT analysis remains the most granular maturity-driver breakdown in the corpus)
research/01-ai-native-landscape/bcg-ai-at-work-2025.md — 5% of organizations capturing substantial financial gains (n=10,600 workers, 11 countries)
research/01-ai-native-landscape/microsoft-work-trend-index-2026.md — Microsoft’s Frontier/Blocked/Stalled quadrant framework: only 19% of AI-using organizations have high individual capability AND organizational readiness; 50% in “Emergent Zone”; 10% “Blocked Agency” (high capability, low org support) (n=20,000 AI users, May 2026)
research/04-consulting-firms/bcg-generative-ai-adoption.md — BCG’s AI consulting practice ($2.7B revenue, 20% of total), the 10-20-70 rule, and the “widening AI value gap” central thesis (60% fail to achieve material returns)
research/07-adoption-challenges/bcg-widening-ai-value-gap-2025.md — “Build for the Future 2025” study (n=1,250, Sep 2025): 5% future-built / 35% scalers / 60% laggards; 3.6x TSR gap; 70% of AI value from people/org/process, only 30% from technology; gap is widening as leaders invest 2x peers (BCG, MEDIUM credibility, TIER 2)
research/07-adoption-challenges/deloitte-state-of-ai-enterprise-2026.md — 60% employee access, 30% governance readiness, 20% talent readiness; 34% transforming core processes (n=3,235 leaders)
research/07-adoption-challenges/deloitte-tech-trends-2026.md — Agentic AI production funnel: 11% in production, 14% near-ready, 38% piloting, 35% no formal strategy; CIOs reporting to CEO rose from 41% (2015) to 65%; AI share of tech budget projected 8%→13%; agentic failure forecast 40% by 2027 (Gartner cited). TIER 2 (n=500+302+622 US tech leaders, March–July 2025)
research/04-consulting-firms/deloitte-ai-roi-paradox-2025.md — TIER 2 (n=1,854, Europe/ME, 2025): typical AI ROI in 2–4 years vs. 7–12 months for traditional IT; five structural barriers to ROI (intangible benefits, siloed platforms, evolving metrics, human factor, entangled transformation); only 10% currently realizing significant agentic ROI
research/04-consulting-firms/accenture-impact-to-advantage-2026.md — Three-phase model: Siloed AI (2–3 years to measurable value) → Structural AI (enterprise architecture rebuilt; AI shifts from experimental to institutional) → Systemic AI (intelligence embedded end-to-end; only a smaller set of organizations reach this stage). Key constraint data: 70% tech budgets on legacy; only 16% with cloud maturity for agentic capture; 9–18 months foundational work before first banking production use case. (n=3,650 C-suite, Jan 2026, MEDIUM TIER 1)
research/04-consulting-firms/accenture-ai-research-2026.md — 8% AI Front-Runner share; 2.5x revenue-growth advantage vs. peers (n=2,000 executives)
research/04-consulting-firms/accenture-reinvention-generative-ai-2024.md — ⚠️ TIER 4 (Oct–Nov 2023 fieldwork): Reinventor/Optimizer/Incremental taxonomy; Reinventors achieved 15pp higher revenue growth and 5.6pp higher margin — strategic-framing benchmark; use for competitive-positioning arguments not current operational benchmarks
research/04-consulting-firms/idc-ai-research-2026.md — IDC FutureScape Oct 2025 + Spending Guide Aug 2025: only 1% of organizations have reached optimized AI-fueled state; 50%+ remain in early transformation; $1.3T spending trajectory to 2029 at 31.9% CAGR; 15% productivity loss quantified for data-unready organizations; 20% of G1000 face AI governance litigation risk by 2030 (HIGH TIER 1–2)
research/11-education-approaches/mit-cisr-digital-colleagues-enterprise-value-2026.md — MIT CISR (n=132 orgs, Sep 2025): 75% expect 25% revenue-per-employee gain; only 9% have formally redefined roles/metrics — the threshold separating digital-colleague leaders from laggards. Regression analysis (p<.05) confirms workflow redesign + role redefinition + high usage intensity as the three statistically significant value predictors. HIGH / TIER 1.
research/09-ai-adoption-cycle/ai-native-adoption-cycle.md — six-stage practitioner adoption cycle
research/05-analyst-firms/gartner-global-labor-market-survey-2026.md — Gartner GLMS n=12,004 (Q1 2026, TIER 1): only 27% of executives have a comprehensive AI strategy; only 20% believe workforce is AI-ready; “enablement illusion” — seat licenses ≠ proficiency; organizations providing tool access without behavioral training create the appearance of AI readiness without the outcomes
research/05-analyst-firms/gartner-cio-agenda-2026.md — only 17% have deployed AI agents despite 91% increasing GenAI budgets; 64% plan agentic AI within 24 months; 18% who reprioritize mid-cycle are 24% more likely to be top performers (Gartner CIO Survey, n=2,501, Jan 2026) (Exploration → Experimentation → Standardization → Scaling → Optimization → Transformation); most enterprises are Stage 2→3; “two-thirds cannot escape pilot purgatory” (McKinsey estimate); stage-transition indicators derived from GitHub/Stack Overflow/JetBrains developer surveys and McKinsey Nov 2025
research/04-consulting-firms/pwc-ai-performance-study-2026.md — 74% of AI economic value captured by top 20% of organizations; top 20% generate ~7.2x more AI-driven revenue/efficiency than average; 60 AI management practices analyzed via PwC AI fitness index; strongest differentiator is growth/reinvention orientation, not efficiency focus (PwC, n=1,217, 25 sectors, Apr 2026)
research/04-consulting-firms/ibm-ibv-dynamic-finance-2026.md — 12% “advanced” finance-function maturity cohort
research/08-radical-vs-tablestakes/spectrum-analysis.md — maps AI engineering maturity from table stakes (code autocomplete) through emerging standard (test generation, AI code review) to radical frontier (autonomous software engineers, self-healing systems); the 2026 table-stakes floor and the pace of category collapse (combining strategic influence + digital agility), empirically derived via K-means clustering and MANOVA validation; cohort reports +37% strategy execution, +19% funding decision speed, +21% ERP ROI, +10% EPM ROI vs. all other segments (IBM IBV + Oracle, n=600 senior finance leaders, Q4 2025)
research/02-corporate-tools/microsoft-work-trend-index-2026.md — Microsoft/Edelman Work Trend Index 2026 (n=16,971 knowledge workers, Feb–Apr 2026): four-quadrant organizational readiness model (Frontier 19% / Stalled 16% / Blocked Agency 10% / Emergent 50%); only 19% reach “Frontier” zone combining high individual capability with high organizational readiness; 67% of AI impact attributable to org factors vs. 32% individual effort (Microsoft TIER 2, vendor-interest caution on Copilot-specific figures).
research/09-ai-adoption-cycle/mid-market-ai-strategy-document.md — Mid-market AI strategy survey: 79% claim to have an AI strategy; only 37% are well-formulated; 42% abandoned the majority of AI initiatives in 2025 — the strategy-to-execution gap as a maturity signal

Gartner CEO Survey 2026: The Task-to-Operational Threshold

Gartner’s CEO and Senior Business Executive Survey (n=469, April 23, 2026) provides the clearest CEO-level data on where organizations sit and where they expect to move:

54% of organizations are currently at task-level automation — discrete AI deployments that do not change process architecture or accountability structure.
Only 13% expect to remain at that level by end of 2028. The other 41% plan to have crossed to operational-level AI: 32% expect self-learning tools assisting human decisions, 27% expect primarily autonomous operations.
This 54%→13% projection aligns precisely with the Stage 2→3 transition in MIT CISR’s four-stage model: both describe the move from demonstrated AI value in pilots to industrialized AI embedded in how the enterprise operates.

Credibility note: Gartner has commercial interest in AI advisory services. Findings are directionally reliable; methodology is disclosed.

Source: research/05-analyst-firms/gartner-ceo-survey-autonomous-business-2026.md

Practitioner voices (pillar 13)

Derek Waldron, Chief Analytics Officer, JPMorgan Chase (Beyond the Pilot / VentureBeat, Apr 13, 2026): “The actual long-term bottleneck for driving maximum value from this technology was not going to be about the model. It was going to be about how the technology connects into the technology estate and data and process estate of an enterprise.” Source: research/13-multimodal-sources/beyond-the-pilot/2026-04-13-what-30k-jpmorgan-ai-agents-taught-me.md
Derek Waldron, Chief Analytics Officer, JPMorgan Chase (Beyond the Pilot / VentureBeat, Apr 13, 2026): “Businesses run on long processes that cross multiple different types of teams. If we want to be able to really move the needle on those processes, there has to be a strategic element to actually rethink what the process itself will need to look like in a world of AI and AI agents.” This is the Stage 2→3 transition in a practitioner’s words: the move isn’t tool selection, it’s process redesign. Source: same.
Vishnu Ram, VP of Engineering, Credit Karma (Beyond the Pilot / VentureBeat, Sep 17, 2025): “We are running something like 65 billion model predictions daily.” Credit Karma at Stage 3 looks like industrial-scale inference integrated into product surfaces — not pilots. Source: research/13-multimodal-sources/beyond-the-pilot/2025-09-17-venturebeat-in-conversation-credit-karmas-path-to-scalable-a.md
Mariana Tessel, EVP and GM, Intuit (Beyond the Pilot / VentureBeat, Apr 2026): “We have seen 3 million customers that are using our agents and Intuit Intelligence. We actually see 85% re-engage work.” Intuit at Stage 3 means AI embedded in the product its customers buy — not an internal productivity tool, but a revenue-generating capability tied directly to retention. Source: research/13-multimodal-sources/beyond-the-pilot/2026-04-01-100m-agents-scaling-the-new-execution-stack-with-intuit.md

What this means for mid-market buyers

Locate your company on the maturity arc before you commit new AI spend. A 300-person company running one proven pilot with measurable value is Stage 2. Adding three more pilots in isolated teams is not progress; it is Stage 2 repetition. The signal of Stage 3 is structural: reusable architecture, AI-redesigned roles, dashboards that track outcomes rather than activity.
The financial break between stages is binary, not gradual. Stage 2 and Stage 3 are separated by roughly 14.8 points of growth and 10.9 points of profit, versus industry average. Moving from Stage 3 to Stage 4 adds only another 5.8 points of growth. The biggest single move is the Stage 2→3 jump — so put resources there, not into chasing Stage 4 capabilities.
Stage is contagious across the organization. A Stage 3 company running Stage 1 in one business unit inherits Stage 1 exposure there (governance gaps, proliferating shadow AI, value leakage). Map stage by function, not just enterprise-wide, and resource the weakest unit up rather than letting the average conceal the exposure.

MIT CISR “Enterprise IT Operating Models in the AI Era” (Thorogood & Woerner, Dec 18, 2025)

A second MIT CISR research thread complementary to the maturity model: rather than describing where companies are on the AI journey, it describes what IT operating model they need to be in to capture cash-flow-positive AI gains.
Four archetypes — Legacy Modernizer (33%), Creative Sprinter (16%), Adaptive Innovator (33%), Efficient Builder (18%) — derived from 39 interviews / 30 companies plus n=152 cross-tab survey.
Right-side archetypes (high IT leadership intensity) capture substantially higher enterprise platform reuse and innovation revenue: Adaptive Innovator at 80% / 71%, Efficient Builder at 65% / 54%, vs. Legacy Modernizer at 53% / 37%.
Frames a candidate explanation for the maturity-model finding that 36% of companies remain stuck below the financial threshold despite multiple years of AI investment: the operating model is the constraint, not the technology.
Headline: “innovation is more effective with strong IT leadership, resulting in high modularity and reuse regardless of whether those enterprises operate in environments with low or high innovation velocity.”

Source: research/01-ai-native-landscape/mit-cisr-enterprise-it-operating-models-2026.md

McKinsey “State of AI Trust in 2026” — Responsible AI Maturity Benchmark (Mar 25, 2026, n=~500)

Adds a governance-specific maturity dimension absent from most AI maturity frameworks. McKinsey’s five-dimension RAI maturity scale (strategy, risk management, data and technology, governance, agentic AI governance) measures how well the organization manages its AI program, not just whether AI is deployed.
Average RAI maturity score: 2.3/4.0 in 2026, up from 2.0 in 2025. Only ~30% reach level 3 in the governance and agentic AI governance dimensions — meaning these lag behind technical capabilities, which is the inverse of what safe agentic deployment requires.
Explicit maturity-to-financial-outcome correlation: organizations investing $25M+ in RAI are significantly more likely to achieve EBIT impact above 5%. This pairs with McKinsey’s earlier State of AI (6% high performers with >5% EBIT) and connects governance investment directly to the financial outcome metric those prior numbers describe.
The accountability shortcut: organizations with named RAI ownership average 2.6 maturity; those without average 1.8. This is the fastest lever for moving up the maturity scale.
Converges with the broader maturity-model pattern: most organizations are at a middle stage (2.3 average) with no clear path to the level (3+) where the framework produces systematized results rather than ad hoc governance.

Source: research/04-consulting-firms/mckinsey-ai-trust-maturity-2026.md

Lenovo/IDC CIO Playbook 2026 (n=3,120) — Late-Stage Adoption Without Governance

The largest IDC-fielded CIO readiness study provides a concrete maturity distribution for enterprise AI in September–October 2025:

60% of organizations are in late-stage AI adoption — the majority of enterprises have moved past experimentation into production.
Only 27% have a comprehensive AI governance framework — meaning most late-stage deployers are at Stage 2 (deployed, not governed) rather than Stage 3 (governed, integrated, outcome-measured).
60% are more than 12 months away from scaling agentic AI — the next wave of the maturity arc is arriving before most organizations are positioned to manage it.
The 27% governance figure directly corroborates the MIT CISR Stage 2→3 bottleneck: most enterprises have demonstrated AI value but have not built the governance, data, and workflow infrastructure to capture it at scale.

Source: research/05-analyst-firms/lenovo-idc-cio-playbook-2026.md — MEDIUM / TIER 2 (Lenovo vendor; IDC independent fieldwork; n=3,120; Sep–Oct 2025)

Forrester AIQ Framework — Organizational AI Aptitude (Apr 2, 2026, n=1,500)

Forrester introduces AIQ (Artificial Intelligence Quotient) as an organizational diagnostic — distinct from individual skill scores. Low AIQ is the mechanism behind Forrester’s own ROI finding (only 13-15% of organizations report EBITDA impact): the technology is deployed; the organizational aptitude to use it correctly is not.
Four behaviors that distinguish high adopters from low adopters form a sequence: (1) define business outcomes and success metrics first; (2) identify use cases aligned to those outcomes; (3) establish a structured deployment runway; (4) scale with cloud, frontier models, and embedded agents. Most organizations attempt step four without completing steps one through three.
The sharpest segmentation data comes from hiring: 54% of high adopters require demonstrated AI skills in hiring vs. 29% of low adopters — a 25-point gap that compounds with every annual cohort.
Data/consulting partnerships separate the cohorts by 21 points (47% high adopters vs. 26% low adopters), suggesting high adopters treat AI value extraction as a capability-building program, not a software deployment.
Connects the AIQ framework directly to existing maturity-model evidence: low AIQ = stuck at Stage 2, unable to make the structural changes (workflow redesign, role redesign, outcome measurement) the Stage 2→3 transition requires.

Source: research/04-consulting-firms/forrester-accelerate-ai-voyage-2026.md

KPMG Global AI Pulse (n=2,110, March 2026) — The Talent Multiplier

KPMG’s largest-ever AI C-suite survey adds a critical multiplier to the maturity model: organizations that invest in talent alongside AI are 4x more likely to report meaningful value (77% vs. 20%). This directly maps onto the Stage 2→3 transition — the cohort stuck at Stage 2 is the one that deployed tools without parallel workforce investment.

74% of global leaders now classify AI as non-discretionary infrastructure, even in a recession scenario
64% report meaningful AI value (AI leaders: 82%; laggards: 20%) — an 18-point spread that mirrors BCG’s Trailblazer/Follower performance gap
32% are already deploying and scaling AI agents; 27% orchestrating multi-agent systems — agent adoption is outpacing most public estimates
Only 20% of early-stage organizations feel confident managing AI risks vs. 49% of AI leaders — governance confidence is a lagging indicator that follows maturity, not a leading one

Source: research/07-adoption-challenges/kpmg-global-ai-pulse-2026.md

Gartner AI Maturity Enterprise 2025 — Longevity and I&O Data

Gartner’s parallel maturity work provides operational longevity and infrastructure metrics not captured in business-outcome frameworks:

High-maturity AI organizations are 2.25x more likely to keep AI projects in production for 3+ years (45% vs. 20%) — longevity, not launch velocity, is the differentiating capability
91% of high-maturity organizations have appointed a dedicated AI leader vs. significantly lower rates at low-maturity peers
Only 39% of technology leaders are confident their AI investments will produce positive financial impact — the confidence gap is wider than the deployment gap
Only 28% of I&O AI deployments fully succeed (Gartner I&O survey, n=782) — most pilots are abandoned, deferred, or delivering below expectations

Source: research/05-analyst-firms/gartner-ai-maturity-enterprise-2025.md

DORA / Google Cloud AI ROI Framework (n=~5,000, Apr 2026)

AI coding tools deliver 35–40% productivity gains on greenfield code; on legacy systems (the majority of enterprise engineering work) the gain drops to roughly 10% or less — the single most important number for any organization buying AI coding tools.
Organizations experience a J-Curve productivity dip during adoption that most defund prematurely; those that sustain through the dip capture the 39% first-year ROI.
“AI magnifies the strengths of high-performing organisations and the dysfunctions of struggling ones” — DORA’s central finding, consistent with every other maturity-correlated study in the corpus.
Maturity-model implication: the legacy-code ceiling is a Stage 2→3 transition signal. Immature organizations, where technical debt is highest, capture the least AI productivity gain and interpret the shortfall as AI failure rather than a workflow-readiness gap.

Source: research/02-corporate-tools/dora-roi-ai-assisted-software-development-2026.md · Google DORA · n=~5,000 · Apr 2026 · MEDIUM-HIGH · TIER 1

KPMG Global AI Pulse Q1 2026 — Risk-Confidence Maturity Bifurcation

Provides a risk-confidence lens on maturity: 20% of organizations in the experimentation phase feel confident managing AI risks vs. 49% of organizations actively scaling agentic AI — a 29-point gap that marks two distinct maturity states, not two points on a single curve.
The gap is built, not declared: scaling organizations have constructed agent-specific identity controls, audit trail infrastructure, blast-radius containment, and human validation workflows. Confidence is a lagging indicator of governance infrastructure, not a self-assessment.
Human validation requirements nearly tripled: 63% of US leaders now require human review of agentic outputs (Q1 2026), up from 22% in Q1 2025 — the fastest-moving maturity signal in the 2026 corpus.
54% of US organizations ($1B+ revenue) are actively deploying agents across core operations as of Q1 2026, up from 11% in early 2024. Deployment velocity has outrun governance readiness across most of the market, which is the structural condition that keeps most organizations at Stage 2.
Methodology: n=237 US C-suite/business leaders ($1B+ orgs); n=2,110 global across 20 markets; fieldwork Feb 17–Mar 17, 2026. KPMG consulting vendor caveat applies.

Source: research/04-consulting-firms/kpmg-global-ai-pulse-tech-report-2026.md

Deloitte “State of AI in the Enterprise 2026” (“From Ambition to Activation”, n=3,235, Aug–Sep 2025)

The most direct empirical split of the maturity distribution at scale: 34% of organizations are genuinely reimagining their business through AI; 30% are redesigning key processes; 37% are using AI at the surface level with minimal process change. These three groups map closely to the Stage 3 / Stage 2 / Stage 1 maturity bands — with the largest cohort (37%) still at the surface level.
Only 25% of organizations currently have 40% or more of AI pilots in production. Fifty-four percent expect to cross that threshold within three to six months. The pilot-to-production transition is the Stage 1→2 inflection point; most of the global enterprise population is still passing through it.
The maturity-to-revenue gap quantifies what Stage 2 delivers without Stage 3: 74% of organizations expect AI to drive revenue growth; only 20% report it today. Efficiency and productivity gains are consistent at Stage 2; revenue impact requires the workflow and business-model redesign of Stage 3.
84% of organizations have not redesigned jobs around AI — the structural absence that keeps organizations at Stage 2. Maturity does not advance through tool proliferation; it advances through the 70% people-and-process investment that the 10/20/70 rule describes.

KPMG Global Tech Report 2026 — Governance Fragmentation as the Maturity Separator (n=2,500, Mar 2026)

The report’s most operationally useful maturity finding is not a headline adoption number — it is the governance-fragmentation differential between high performers and average organizations.

Only 2% of high performers report “several disconnected AI projects and teams” versus 34% of average performers — a 17x gap that is the primary structural explanation for the 4.5x ROI differential (high performers) versus 2x ROI (industry average).
The fragmentation gap maps directly onto the Stage 2→3 transition: organizations with distributed AI projects are producing pilots-without-scale; organizations with centralized governance are compounding returns across the enterprise.
Three governance decisions separate high performers: centralized investment prioritization (91%), centralized technology/supplier selection (85%), centralized talent strategy (82%).
Tech debt compounds the effect: 8% of high performers say tech debt prevents new AI investments versus 45% of average performers — confirming MIT CISR’s finding that the Stage 2→3 transition requires infrastructure resolution, not just more AI tools.
74% of tech executives report AI delivering business value; only 24% achieve ROI across multiple use cases. The 50-point gap has worsened (−7pp from prior survey period), suggesting the ambition-execution divergence is widening as AI deployment scales without governance scaling.

Source: research/04-consulting-firms/kpmg-global-tech-report-2026.md

Gartner Autonomous Business Layoffs Study (n=350, Q3 2025, May 2026)

A maturity-relevant finding on the wrong sequencing: 80% of autonomous AI deployers reduced workforce, yet workforce reduction rates were nearly equal across high-ROI and low-ROI organizations. Maturity advancement requires human capability investment, not just headcount optimization.
High-ROI organizations are distinguished by investment in skills (retraining to direct autonomous systems), new roles (workflow architects, AI oversight), and operating model redesign — the same three dimensions that define the Stage 2→3 transition in MIT CISR’s framework.
This is structural confirmation that AI maturity cannot be purchased by reducing the workforce. The Stage 3 differentiator is organizational capability, not headcount delta.

Source: research/05-analyst-firms/gartner-autonomous-business-layoffs-roi-2026.md

Cross-references with BCG’s AI-First Cost Advantage (3x cost reduction for AI leaders vs. peers), McKinsey’s 6% high-performer cohort findings, and Forrester’s AIQ segmentation all point to the same Stage 3 structural gap: organizations that capture disproportionate value have redesigned workflows end-to-end, not merely enabled employees with tools.

Source: research/04-consulting-firms/deloitte-state-of-ai-enterprise-2026.md

OpenAI B2B Signals — The Frontier Gap as Maturity Proxy (May 2026)

Frontier firms (95th percentile AI use intensity) now consume 3.5x as much AI intelligence per worker as typical firms, up from 2x one year ago. Volume accounts for only 36% of this gap; 64% is depth — more complex tasks, richer context, agentic workflows.
The clearest single maturity signal is 16x agentic (Codex) message volume at frontier firms vs. typical firms. Seat count and message volume are Stage 1 metrics; agentic task delegation is a Stage 3 signal.
The frontier gap is widening — which is the B2B Signals empirical confirmation of the maturity-stage financial performance spread MIT CISR measures: organizations that made the Stage 2→3 structural shift are accelerating away from those still replicating Stage 2 behaviors.
Vendor caveat: OpenAI is a commercial vendor; population is active OpenAI enterprise customers, not a representative enterprise cross-section. Directional finding consistent with BCG and McKinsey data.

Source: research/01-ai-native-landscape/openai-b2b-signals-enterprise-ai-depth-2026.md

Gartner I&O AI Deployment — Only 28% Fully Succeed (n=782, Nov–Dec 2025)

Only 28% of AI use cases in infrastructure and operations fully succeed and meet ROI; 20% fail outright. The 57% of I&O leaders reporting at least one failure overwhelmingly cite over-ambition — expecting too much, too fast — as the primary cause. This maps directly to Stage 1→2 transition failures: deploying AI into complex, unpredictable environments before proving value in bounded, measurable ones.
53% of I&O AI wins occur in ITSM (ticket classification, knowledge retrieval, automated responses) — the most mature, bounded I&O use cases. Failures concentrate in auto-remediation and agent-led cross-system workflows, the most ambiguous high-autonomy deployments. Use-case maturity selection is the Stage 2 discipline; skipping it produces the 20% failure rate.
Two blockers appear in equal measure among I&O failures: skills gaps (38%) and data quality/availability (38%). Neither is a technology problem. This corroborates the MIT CISR finding that Stage 2 organizations are distinguished by workforce capability investment and data foundation readiness — not by model sophistication.

Source: research/05-analyst-firms/gartner-ai-io-stalling-2026.md

Grant Thornton “2026 AI Impact Survey” — Integration-Stage Revenue Divide (n=950, US, Feb–Mar 2026)

Provides the sharpest revenue-growth maturity split in the 2026 corpus: 58% of fully integrated organizations report revenue growth vs. 15% still piloting — a 43-point gap consistent with but larger than the McKinsey 6% / BCG 5% high-performer segmentations, which use different integration criteria.
Governance confidence tracks integration stage tightly: 74% of fully integrated organizations are “very confident” in their governance readiness; only 7% of piloting organizations say the same. This is the clearest empirical linkage between governance maturity and integration-stage advancement.
78% of all organizations lack confidence they could pass an independent AI governance audit in 90 days — the single most actionable maturity benchmark in the corpus because it translates directly to an internal diagnostic the board can demand Monday morning.
Apply caveat: self-reported integration stage; US-only; operations-heavy sample (41%); revenue figures are respondent-attributed, not measured against a control group.

Source: research/04-consulting-firms/grant-thornton-ai-impact-survey-2026.md

Failure-Mode Synthesis — What Structural Pre-Conditions Predict Failure at Scale

A corpus synthesis (April 2026) drawing from McKinsey (n=1,993), Grant Thornton (n=950), Writer (n=2,400), OutSystems (n=~1,900), MIT CISR FinCo case, METR RCT (n=16), and Atlan 200-deployment analysis identifies four pre-conditions that appear in nearly every documented failure:

No workflow redesign mandate — tool deployed into unchanged workflow; no cross-functional authority to eliminate steps.
No named governance owner — McKinsey RAI maturity benchmark: named owner = 2.6 average score vs. 1.8 without. FinCo built comprehensive governance apparatus with no single owner and ended up with more shadow AI than before.
No data readiness assessment — Gartner predicts 60% of AI projects abandoned through 2026 for this reason. Piloting on curated sample data is not production validation.
No production path in the pilot design — 46% of POCs scrapped (S&P Global, n=1,006). Pilot approved as experiment; production requires different budget authority that was never secured.

The survivorship caveat: every public success case in the corpus (Palantir AIPCon, Vodafone, SlickDeals, TELUS) represents a deployment where the organization agreed to publish. The 85-95% that did not reach comparable outcomes are structurally absent from public research.

Source: research/07-adoption-challenges/ai-deployment-failure-modes-synthesis.md

Genpact / HFS Research — The Readiness-Expectation Gap (n=545 Fortune 2000, 2026)

92% of Fortune 2000 senior executives believe agentic AI will fundamentally change operations; only 13% have it integrated today. The 79-point gap is the clearest quantification of the maturity-expectation chasm available in 2026 data.
Primary barrier is not technical: regulatory exposure, reputational risk, lack of explainability, and unclear accountability structures are the reported obstacles — the governance failure modes that stall Stage 2→3 transitions.
Skills demand signal: workflow orchestration (42%), data engineering (39%), and monitoring/observability (36%) are cited as top needs — operational roles, not AI engineering. This matches the Stage 3 capability profile in which integration competency outweighs model competency.

Source: research/01-ai-native-landscape/genpact-hfs-autonomy-requires-trust-agentic-ai-2026.md

ServiceNow / Oxford Economics AI Maturity Index 2026 — Second Consecutive Decline (n=4,473, Apr 2026)

The most counterintuitive single maturity finding in the 2026 corpus: enterprise AI maturity scores fell for the second consecutive year, even as AI investment rose.

Average score: 35/100 — down from 44 in 2025, which itself was down from the 2024 baseline. A 20% year-over-year decline.
Fewer than 1% of organizations scored above 50. The highest individual score fell 13 points from the prior year.
AI Pacesetters (18.2% of respondents) are not immune: their average fell from 54 to 44.
The explanation is structural, not cyclical: organizations rated against generative AI copilot benchmarks in 2025 are now assessed against agentic AI standards they have not yet met. The bar moved; the organizations did not.
The widening performance gap is measurable on exactly the two dimensions that define Stage 2→3 transition: workflow redesign (54% Pacesetters vs. 12% rest) and data integration (60% vs. 41%).

ServiceNow has direct commercial interest in enterprise AI platform adoption; Oxford Economics conducted independent fieldwork. The counter-to-vendor-interest finding (maturity declining despite rising investment) adds credibility. Corroborates McKinsey’s 6% high performer cohort and MIT CISR’s Stage 2→3 bottleneck from an independent instrument.

Source: research/05-analyst-firms/servicenow-enterprise-ai-maturity-index-2026.md — MEDIUM / TIER 1 (Oxford Economics independent fieldwork, n=4,473, Apr 2026)

Post-Deployment Lifecycle Management (April 2026)

Maturity stage describes where a company sits at a point in time. Lifecycle management describes what it costs to stay there — and why the majority of organizations regress from Stage 2 to effectively Stage 1 output quality within 12-18 months of their first successful deployment.

The core finding: 91% of ML models degrade over time (MIT research, 32 datasets across 4 industries). LLM-based deployments do not degrade in the traditional sense, but vendor model updates — which are frequent and sometimes breaking — require ongoing prompt engineering maintenance that most organizations do not budget for. Models unchanged for six months see error rates jump 35% on new data. Organizations without monitoring infrastructure do not discover this until the damage has accumulated.

Four maintenance deficits explain the majority of ROI gaps at 12-18 months post-deployment:

No monitoring budget — 75% of businesses observe performance declines without proper monitoring. Drift accumulates for months before detection.
No prompt engineering maintenance budget — vendor model updates (GPT-4 → GPT-4o → GPT-5) require prompt re-engineering that organizations treat as zero-cost. Annual cost for a mid-market deployment: $20,000-$60,000.
No governance re-audit budget — EU AI Act enforcement (August 2026) requires quarterly reviews for medium-risk systems and monthly audits for high-risk systems. Most organizations treat governance as a one-time pre-deployment activity. 78% of organizations cannot pass an independent AI governance audit in 90 days (Grant Thornton, n=950, Feb-Mar 2026).
No switching cost reserve — the average AI platform migration costs $315,000 (Swfte AI enterprise survey, 2025). Forced migrations from vendor deprecations (OpenAI Assistants API shutdown August 2026, Sora API shutdown September 2026) impose this cost on organizations that did not negotiate exit provisions at contract time.

The maturity-lifecycle intersection: Organizations at MIT CISR Stage 2 that experience deployment quality degradation without monitoring often interpret the decline as evidence that “AI doesn’t work” — misdiagnosing a lifecycle management failure as a technology failure. This triggers abandonment rather than remediation. The 42% of organizations that abandoned at least one AI initiative in 2025 (Deloitte, n=3,235) are partially explained by this pattern: successful early deployments whose quality was not maintained.

Annual lifecycle maintenance budget for a 500-person company with a mature Stage 2 deployment across 4-6 workflows: $110,000-$305,000 per year, on top of licensing and usage costs. This figure is absent from most business cases and ROI projections.

Source: research/07-adoption-challenges/ai-deployment-lifecycle-tco-management.md

Competitive Monitoring: How to Know Where Peers Sit on the Maturity Arc

Maturity models are useful for self-assessment. They are equally useful for reading competitor posture — if you know what signals correlate with each maturity stage.

Stage 1 (experimenting): No AI/ML roles in job postings. No AI operational metrics in earnings calls. No vendor case study names. Possibly: one “AI strategy” executive hire.

Stage 2 (scaling pilots): AI/ML engineering roles appearing in postings. Earnings call language is aspiration-forward (“investing in AI,” “excited about our partnership”). Vendor pilot announced but no outcome metrics.

Stage 3 (industrialized): Surge in data engineering hires (Databricks, dbt, Snowflake skills). Earnings calls cite operational outcomes: cycle-time reductions, headcount reallocations, throughput increases. Named in vendor case studies with quantified results. Product announcements include AI-specific feature differentiation.

The Federal Reserve (Apr 2026) finds only 5.5% of firms had posted AI-related jobs by late 2025 — vs. 10% of firms claiming AI use in surveys. The gap between claimed and deployed is the monitoring opportunity: firms posting AI/ML jobs have committed capital and are moving toward Stage 2; firms that have stopped posting and shifted to data engineering have crossed into Stage 3 production.

BCG (Sep 2025, “AI Leaders Outpace Laggards”): Stage 3+ firms (“future-built”) show 1.7x revenue growth, 3.6x three-year TSR, and 1.6x EBIT margin vs. Stage 1-2 firms. The gap is compounding because future-built companies reinvest AI returns into the next capability cycle — spending 26% more on IT and 64% more of that IT budget on AI in 2025.

Source: research/08-radical-vs-tablestakes/ai-competitive-intelligence-monitoring.md

Roland Berger “Industrializer” Segmentation (n=203, Mar 2026)

Independent validation of the 5–10% high-performer pattern from a strategy consulting firm with no AI platform commercial interest.

Four-segment model: Industrializers (~10%), Stalled (high ambition + spend, low returns), Observers (piloting without scaling), and early explorers. Pattern cuts across geographies, industries, and company sizes — the gap is structural, not sectoral.
The Industrializer Code: Retain AI control while using partners; integrate into systems (not wrappers); embed governance into platforms; federate innovation through common standards; treat go-live as the beginning of operational investment, not the end of a project.
What Industrializers don’t do: Spend more or move faster than peers. The differentiator is operational model, not budget or velocity.
The decisive shift: “Technology readiness is no longer the binding limitation, and access to capital is rarely the limiting factor. What holds many organizations back is a lack of engineering discipline in how AI is governed and operated at scale.”
C-suite mandate by function: CEO → orchestrate, not innovate; CTO/CIO → build infrastructure that makes right behavior easy (embed governance in platforms); CFO → fund shared capabilities, not isolated project budgets.

Source: research/04-consulting-firms/roland-berger-profitless-prosperity-ai-2026.md · Mar 2026 · MEDIUM-HIGH · TIER 1

a16z Startup Penetration as Maturity Proxy (April 2026)

a16z’s April 2026 analysis provides a complementary maturity metric: share of enterprises with live, paid contracts with AI-native startups (excluding incumbent platforms).

29% of Fortune 500 have signed, paid, production deployments with AI-native startups — the clearest evidence of Stage 2+ commitment among the largest companies
~19% of Global 2000 have equivalent deployments
The three dominant use cases — coding, customer support, and knowledge search — map precisely to high-maturity workflow conditions: text-based, repetitive, human-review-ready, verifiable

This is a different measurement than survey-based “AI in at least one function” (McKinsey 88%). It measures willingness to sign a budget-line contract with an AI-native vendor alongside incumbents — a stronger commitment signal than pilot participation. Organizations at this stage have crossed the Stage 2 threshold and are testing whether their workflow architecture supports Stage 3 industrialization.

Source: research/01-ai-native-landscape/a16z-enterprise-ai-adoption-where-2026.md

Return on AI Institute: Six-Stage AI Economic Maturity Model (March 2026)

Davenport and Srinivasan (Return on AI Institute, HBR, March 2026, n=1,006 executives, 11 countries, 32 industries) introduce a six-stage AI Economic Maturity Model as a research-based progression roadmap from pilots to high-value returns.

The full stage definitions are paywalled. The key structural insight: the seven factors that separate high-return organizations from low-return organizations are not technology choices — they are operational disciplines. The primary differentiator is “how deliberately you measure, manage, and report the value those tools create,” not which tools are deployed.

90% of organizations in this survey report getting any value from AI (45% “great,” 45% “moderate,” 9% “small”); this is not in conflict with McKinsey’s 6% high-performer finding — they measure different outcome levels
2x value likelihood for organizations with clean, complete, and current data vs. fragmented sources (Factor 1)
>80% vs. <50% adoption rate for structured change management vs. “set-and-forget” deployments (Factor 4)
−15% effectiveness decline within 6 months for static deployments without monitoring/retraining (Factor 6)
71% of global CIOs say AI budgets will freeze or be cut if value cannot be demonstrated within two years

The maturity model is consistent with: MIT CISR 4-stage (Stage 2→3 transition as the financial break point), BCG (5% “substantial gains”), McKinsey (6% EBIT-impact high performers). All converge on a pattern where a small minority advances through the full progression and compounds returns; the majority stalls at early stages without structural operational change.

Source: research/04-consulting-firms/davenport-return-on-ai-institute-hbr-2026.md

Board-Ready AI Strategy Briefing — Director AI Literacy Gap (Deloitte/NACD/Harvard Law, 2024–2025)

79% of board members report limited, minimal, or no AI knowledge — yet 88% of their organizations are deploying AI in at least one function. The gap is the largest governance risk in American mid-market companies. (Deloitte, n=468, 57 countries, May–July 2024; TIER 3 — flag date when citing)
Only 28% of S&P 100 companies disclose both board-level AI oversight and a formal AI policy. Proxy advisors are moving toward withhold recommendations for directors unable to demonstrate AI literacy.
State AI laws (Colorado, Texas, California), SEC enforcement against “AI washing,” and enterprise client due diligence questionnaires all require documented board engagement with AI strategy and risk.
A 200–500 person company needs 5–8 slides for a board AI briefing, not 50. Directors want narrative answering: What are we doing? What is the risk? What next?
Maturity implication: organizations where the board cannot articulate AI strategy are structurally blocked from Stage 3 advancement — fiduciary pressure is now an external forcing function on maturity progression.

Source: research/09-ai-adoption-cycle/board-ready-ai-strategy-briefing.md

MIT CISR — AI-Savvy Boards Drive Superior Performance (Dec 2025, n=2,800)

Only 26% of large U.S. company boards qualify as “AI-savvy” under MIT CISR’s updated criteria (3+ directors with hands-on GenAI/agent experience). The remaining 74% govern AI by analogy to prior tech waves — and the performance gap is measurable.
AI-savvy boards outperform industry peers by 10.9 percentage points in ROE; non-savvy boards underperform by 3.8 points. The 14.7-point spread is the sharpest board-level maturity proxy in the corpus.
Market cap premium: AI-savvy boards carry $15.5B above industry average; non-savvy sit $5.4B below — a $20.9B combined spread across 2,800 publicly traded companies.
Sector distribution: Health care (8% AI-savvy), Construction (6%), Retail-automotive (11%) — the industries with the largest governance gaps relative to their AI exposure risk.
Maturity implication: board AI literacy is now a Stage 3→4 gating condition. Organizations where directors cannot evaluate agentic AI capital allocation or governance thresholds cannot advance maturity without board refreshment or structured director education.

Source: research/04-consulting-firms/mit-cisr-ai-savvy-boards-superior-performance-2025.md

NBER w34836 “Firm Data on AI” — Independent Academic Baseline (Feb 2026, n~6,000, US/UK/DE/AU)

69% of firms actively use AI — adoption is no longer a leading indicator of financial impact; it is table stakes
~90% of senior executives report no measurable past impact of AI on either employment or productivity over the prior three years — the largest independent academic confirmation of the maturity-model gap between deployment and outcomes
Executives average only 1.5 hours per week of personal AI use; over two-thirds use AI regularly but at minimal intensity — consistent with the BCG/McKinsey finding that executive engagement is the single most predictive variable of AI program advancement
Same executives predict +1.4% productivity, +0.8% output, -0.7% employment from AI over the next three years — near-zero past impact alongside meaningful future expectations is the perception gap that defines the current Stage 2 majority
The 90% / no-past-impact finding is the cleanest independent confirmation of the pattern across all maturity frameworks: tools deployed, workflows not redesigned, results not yet visible in aggregate numbers

Source: research/01-ai-native-landscape/nber-firm-data-on-ai-executive-perception-2026.md

IBM IBV / Oxford Economics “The Enterprise in 2030” — The Ambition-Architecture Gap (Jan 2026, n=2,007)

79% of executives expect AI to significantly contribute to revenue by 2030 — but only 24% can articulate where that revenue will come from. This 55-point expectation-to-roadmap gap is the clearest empirical statement of the Stage 2 problem: confidence in the destination without the architecture to reach it.
68% worry their AI efforts will fail due to lack of integration with core business activities — naming the same failure mode that the MIT CISR Stage 2→3 transition framework describes: pilots that don’t connect to how the business actually runs.
Organizations that scale AI across multiple workflows (rather than isolated use cases) anticipate 24% greater productivity gains and 55% higher operating margins than peers by 2030 — the integration premium that distinguishes Stage 3 from Stage 2.
AI-first organizations anticipate 70% greater productivity improvement, 74% greater process cycle time reduction, and 67% greater project delivery improvement vs. non-AI-first peers — converging with the BCG 5% / McKinsey 6% finding that the gap between movers and majority is structural, not incremental.
57% say competitive advantage by 2030 will come from AI model sophistication; only 28% have clarity on which models they’ll need — the model clarity gap is the next maturity threshold, emerging as multi-model architectures (expected by 82% of executives) replace single-platform deployments.

Source: research/07-adoption-challenges/ibm-ibv-enterprise-2030-ai-ambition-gap-2026.md

Practitioner Tools: Self-Assessment and Readiness

AI Maturity Peer Benchmarking Methodology — Step-by-step guide for CIOs/CAIOs to benchmark AI maturity against peers using consistent dimensions (workflow coverage, governance depth, measurement infrastructure, executive engagement). Translates MIT CISR and BCG maturity stages into a diagnostic that can be completed in 90 minutes without a consulting engagement. Source: research/07-adoption-challenges/ai-maturity-peer-benchmarking-methodology.md
AI Readiness Scorecard Pre-Deployment — Pre-launch readiness assessment covering data quality, governance structure, change management plan, and measurement baseline. The operational companion to maturity frameworks — answers “are we actually ready to deploy this use case?” before the first workflow goes live. Source: research/07-adoption-challenges/ai-readiness-scorecard-pre-deployment.md
30-Minute AI Workflow Readiness Assessment — 20-question, 5-section assessment that evaluates a specific workflow (not the company) across Data Foundation, Decision Architecture, Human Oversight Design, Adoption Readiness, and Scale Architecture. Operationalizes the Stage 2→3 transition criteria at the workflow level; scoring table maps 0–40 scale to Proceed/Gaps/Not Ready. Source: research/09-ai-adoption-cycle/ai-workflow-readiness-30-minute-assessment.md
AI Deployment Red-Flag Checklist — 18-item pre-deployment checklist derived from corpus failure evidence (McKinsey n=1,993, FinCo HITL failure, OutSystems adoption collapse, METR RCT, Rewired’s five transformation sins). Items are organized by severity; 3+ flags present = stop and remediate before investment. Source: research/09-ai-adoption-cycle/ai-deployment-failure-mode-red-flag-checklist.md
Data Reset Decision Tree — Workflow-level decision tool classifying any planned AI deployment as Proceed (existing architecture), Prepare (2–4 week cleanup), or Reset (full ontology/schema rebuild). Primary variable is domain count. Includes four-workflow-archetype table (transactional/analytical/generative/agentic) with expected path per type. Source: research/09-ai-adoption-cycle/ai-data-reset-decision-tree.md
Rewired Transformation Roadmap Template — 12-month quarterly roadmap keyed to Rewired’s six-capability sequencing. Maps directly to the Stage 2→3 transition: Q1 establishes domain concentration and data architecture classification; Q2 delivers workflow redesign + pilot on production data; Q3 builds reusable data products and governance; Q4 validates P&L and deploys second workflow using reuse infrastructure. Each quarter maps to a specific MIT CISR maturity transition requirement. Practical enough for a PMO. Source: research/09-ai-adoption-cycle/rewired-transformation-roadmap-template.md

Forrester “Accelerate Your AI Voyage” — Still Chasing Transformative Value (Apr 2026, n=1,500)

Only 15% of AI decision-makers report an EBITDA lift in the past 12 months — three years into widespread GenAI adoption. Most enterprises remain at Stage 2: activity without financial proof.
Fewer than one-third can tie AI value to P&L changes — the measurement infrastructure gap that defines Stage 2 organizations in MIT CISR and BCG frameworks.
48% of firms have already cut headcount due to AI; financial return is running well behind workforce reduction — naming the misalignment risk a CFO should flag before authorizing the next deployment.
High adopters are differentiated by customer-facing use case focus, CEO-driven strategy, data infrastructure investment, and structured talent development — the same four Stage 3 variables identified across maturity frameworks.
Low AI fluency (“AIQ”), siloed function-level adoption, and overemphasis on marginal productivity use cases are the three primary barriers named — consistent with MIT CISR’s “Experimentation” stage description.

Source: research/05-analyst-firms/forrester-genai-enterprise-value-2026.md

Data & AI Leadership Exchange 2026 — Fortune 1000 Production Benchmark (n=~110 CDOs/CAIOs)

The Randy Bean / Davenport annual benchmark provides a longitudinal view of AI production adoption at Fortune 1000 level — the clearest data on what the top of the maturity distribution looks like in 2026:

93.6% have AI in production (limited + at scale) — up from 29.2% just two years ago. The Fortune 1000 experimentation era is over.
39.1% in production at scale — up from 4.7% two years ago (8x increase). Scale deployment, not just limited production, is now the majority direction.
54% report high or significant business value — up from 47.6% last year. Even at this elite cohort (dedicated CDOs, multi-year programs, production deployments), 46% still do not report high/significant value.
93.2% cite culture/change management as #1 barrier — the highest in 15 years, even among the most AI-advanced companies. Technology readiness is solved; organizational readiness is not.
The 54% high/significant value figure at Fortune 1000 CDO level corresponds roughly to MIT CISR Stage 3+, BCG “trailblazers,” McKinsey’s 6% EBIT-impact cohort, and BCG’s 5% substantial-gains cohort — all describing the same minority cohort that has crossed the production-to-value threshold.

Source: research/04-consulting-firms/randy-bean-ai-data-leadership-benchmark-2026.md

Capgemini AI Perspectives 2026 — The Operationalization Plateau (Jan 2026, n=1,505)

38% have operationalized generative AI beyond pilots — consistent with the BCG/McKinsey/Deloitte pattern and measuring a different threshold (at least one use case in production, not necessarily with financial impact). The 38% brackets the McKinsey/BCG 5–8% financial performance cohort: most organizations that have deployed AI are not yet capturing measurable financial returns.
63% are pruning low-value AI projects — behavioral evidence of the transition from Stage 1 (experimentation) to Stage 2 (selection): organizations are no longer adding pilots indiscriminately; they are concentrating on fewer, higher-impact deployments.
Top enablers of Stage 2→3 transition: executive sponsorship (67%), workforce upskilling (60%), governance frameworks (53%), data infrastructure (51%) — same four levers identified in BCG AI at Work, MIT CISR maturity research, and Accenture Front-Runners studies.
5% of budget on AI in 2026 (up from 3% in 2025), with >50% committing to 5-year horizons — the budget time horizon shift from annual ROI to multi-year infrastructure investment is the clearest marker of organizations moving toward Stage 3.
China geographic gap: ~50% of Chinese organizations piloting/deploying agentic AI vs. lower rates in US/Europe — a leading indicator of where the next operationalization gap will emerge.

Source: research/04-consulting-firms/capgemini-ai-perspectives-2026.md

Oliver Wyman Forum CEO Agenda 2026 — The Deployment-Leader Performance Gap (Apr 2026, n=415 CEOs)

The only large-n CEO-specific AI ROI survey in the corpus (most comparable studies survey senior executives broadly). The maturity signal: the deployment-leader / laggard performance gap is wider here than in any other dataset because the n is CEOs — organizational design decisions, not tool purchase decisions.

27% overall CEO ROI satisfaction (met or exceeded expectations) — down from 38% a year ago, but the aggregate masks a structural split by maturity tier.
49% of deployment leaders (CEOs scaling across 2+ business categories = the corpus’s clearest CEO-level Stage 3 proxy) say ROI met or exceeded expectations, versus 17% of laggards — a 32-point performance gap at CEO level.
Workflow redesign rate: 49% (leaders) vs. 38% (average) — the 11-point gap in a CEO survey signals that the maturity differentiator is operating at the organizational design level, not the tool selection level. Consistent with BCG 10/20/70 rule and McKinsey’s human capital primacy finding.
53% say it is too early to assess (up from 41%) — the largest single cohort. This is not skepticism; it is a capital-deployment-ahead-of-measurement-systems pattern that characterizes organizations in the transition between Stage 2 and Stage 3.
~25% report zero revenue impact — concentrated in companies that deployed onto existing workflows without redesigning them. The zero-impact cohort is the Stage 2 → Stage 3 bottleneck population.

Source: research/04-consulting-firms/oliver-wyman-ceo-agenda-2026.md

Foundry State of the CIO + AI Priorities Study 2026 — The Budget-Without-Strategy Maturity Trap (n=911 + n=538, 2026)

The Foundry dual survey provides the clearest CIO-specific evidence of what Stage 1→2 stall looks like in budget terms: organizations funding AI without the governance structure to convert that funding into returns.

97% piloting or implementing AI — near-universal implementation signals that the deployment threshold no longer differentiates organizations. Maturity stage is now determined by what happens after initial deployment.
65% have a dedicated AI budget (up from 36% two years ago) — the budget formalization is a Stage 2 marker. Organizations that have not formalized dedicated budgets are almost certainly still in Stage 1 experimentation.
31% still lack a clear corporate AI strategy despite the budget growth — this is the Stage 1→2 transition failure pattern. Budget without strategy produces the 97%/6% paradox (near-universal deployment, McKinsey 6% EBIT impact).
33% cannot determine ROI — measurement capability is the Stage 2 floor. Without it, organizations cannot distinguish progress from activity, cannot justify continued investment at the board level, and cannot identify which deployments to scale vs. exit.
Top investment priorities are now infrastructure, skills, and measurement — the Stage 2→3 investment pattern. The organizations reaching this investment sequence are the ones executing the BCG 10/20/70 rule in practice.

Source: research/05-analyst-firms/foundry-state-of-cio-ai-priorities-2026.md · 2026 · MEDIUM-HIGH · TIER 1

McKinsey State of Organizations 2026 — The 88/81 Maturity Diagnostic (n=10,018, Jun–Sep 2025)

The largest executive sample in the corpus (n=10,018, 15 countries, 16 industries) provides the sharpest quantification of the Stage 2 plateau.

88% deploying AI, 81% reporting no meaningful bottom-line impact. This 7-point gap is the Stage 2 plateau at population scale. The organizations in the 81% are not failing to deploy — they are failing to convert deployment into EBIT.
Only 1% describe their gen AI rollout as “mature.” Only 6% are realizing full value from advanced technologies. Maturity is not a function of time — it is a function of governance design, workflow architecture, and people investment.
The AI Pioneer cohort (23% of respondents) is the closest equivalent to a Stage 3 proxy in this dataset. Pioneer differentiators: 90% of leaders actively champion AI (vs. 14% average); 56% of Pioneers believe employees will achieve more (vs. 26% non-Pioneers); AI rolled out across most departments.
Governance accountability is the Stage 2→3 gating mechanism. 1 in 6 organizations has no clear C-suite AI owner. High performers are 3x more likely to report that senior leaders demonstrate explicit ownership of AI outcomes. The maturity bottleneck is not capability — it is named accountability.
5:1 investment ratio is the resource allocation diagnosis. Organizations that sustain top-tier performance are 4x more likely to prioritize people investment alongside technology investment. Stage 2 organizations typically have an inverted ratio — technology investment dominates and workflow/people change is underfunded.

Source: research/04-consulting-firms/mckinsey-state-of-organizations-2026.md

ServiceNow / Oxford Economics Enterprise AI Maturity Index 2026 — The Declining-Maturity Signal (n=4,473, Apr 2026)

The second annual edition of this index provides a longitudinal data point absent from every other maturity model in the corpus: average maturity is declining despite rising AI investment.

Average score fell from 44 to 35 (−9 points, −20%) year-over-year on a 100-point scale across five pillars. This is the second consecutive annual decline since the index launched in 2024.
Fewer than 1% of respondents scored above 50. The highest individual score fell 13 points from the prior year. No organization in the study achieved what the index defines as high maturity.
Pacesetters (18.2% of respondents) are not immune. Their average fell from 54 to 44 — still 9 points above average, but declining at the same rate.
The Pacesetter gap is widening on execution dimensions: 54% of Pacesetters invent new human-AI workflows vs. 12% of others; 60% connect data and remove silos vs. 41% of others. These are the BCG 10/20/70 and MIT CISR workflow-redesign differentiators at scale.
The maturity bar moved, not the organizations. The emergence of agentic AI reset the reference point against which organizations self-assess. Organizations that were “mature” on generative AI copilots are now measured against agentic AI readiness — and most have not closed that gap.
Geographic signal: Europe and Middle East average dropped 10 points YoY, consistent with the broader global decline pattern.

Source: research/05-analyst-firms/servicenow-enterprise-ai-maturity-index-2026.md · Apr 2026 · MEDIUM · TIER 1

Cisco AI Readiness Index 2025 — The Persistent 13% (n=8,039, Aug 2025)

Third annual edition of Cisco’s readiness study — one of the largest AI maturity datasets in the corpus. n=8,039 senior business leaders at 500+ employee organizations across 30 markets. Double-blind survey, independent analysis (Satori Experience). TIER 1. Source credibility: MEDIUM-HIGH (infrastructure vendor commercial interest; consistent with ServiceNow/Logicalis/WalkMe maturity findings).

13% of organizations globally have been Pacesetters for three consecutive years. The denominator is not changing. Organizations that have not achieved Pacesetter status are not simply behind schedule — they are missing a structural approach to all six readiness pillars simultaneously.
87% remain in Chaser, Follower, or Laggard categories despite three years of compounding AI investment. Time and budget are not the gap-closers; system-level discipline across Strategy, Infrastructure, Data, Governance, Talent, and Culture is.
Culture is the weakest pillar: only 9% of all organizations are at Pacesetter level for Culture (vs. 40% for Talent). The largest share (45%) are Followers. Tool deployment consistently outruns cultural readiness.
Data has the most bimodal distribution: 29% Pacesetter vs. 44% Laggard — the lowest Laggard count of any pillar becomes the highest. Organizations either have clean, centralized data or they don’t. No middle ground produces Pacesetter data outcomes.
Pacesetters are 4x more likely to move pilots into production. 77% have finalized use cases vs. 18% of all companies. The production gap between Pacesetters and the rest is larger than the investment gap.

Source: research/07-adoption-challenges/cisco-ai-readiness-index-2025.md · Aug 2025 · MEDIUM-HIGH · TIER 1

Logicalis Global CIO Report 2026 — The Governance Compromise Pattern (n=1,000+, Mar 2026)

The 12th Annual Logicalis Global CIO Survey provides the CIO-perspective complement to maturity models built from executive or analyst data.

Only 33% of CIOs believe their organization can scale AI beyond initial deployments. This is the practitioner self-assessment of Stage 2→3 transition probability — two-thirds of technology leaders executing AI programs do not think their organizations will reach enterprise scale.
62% are already compromising on governance due to limited knowledge. Stage 2 organizations are not skipping governance by choice — they lack the expertise to implement it at the speed deployment demands.
76% call unchecked AI a serious concern; only 44% fully grasp the risks they’re accepting. The gap between concern and understanding is the governance knowledge deficit in quantified form.
89% operate on “learning as we go.” No playbook exists in the median enterprise; governance frameworks lag deployment by 12–18 months.
Skills constraint is the primary maturity bottleneck: ~89% cite skill deficits (not budget or model capability) as what limits AI ambitions.

Source: research/05-analyst-firms/logicalis-cio-report-2026.md · Mar 2026 · MEDIUM · TIER 1

Atlassian State of Teams 2026 — The Workflow Integration Gap (n=12,035, Jan–Feb 2026)

Largest knowledge-worker sample in the 2026 corpus on the gap between AI access and embedded workflow use. n=12,035 knowledge workers + 173 Fortune 1000 executives, double-blind, Jan–Feb 2026. Source credibility: MEDIUM-HIGH (Atlassian commercial interest in enterprise teamwork narrative; consistent with BCG/Deloitte/Stanford maturity findings). TIER 1.

85% of knowledge workers use AI; only 29% have embedded it in actual workflows. The 56-point gap between access and embedded use is the maturity model’s single most important operational measurement — it separates Stage 1 (tool access) from Stage 2 (workflow integration).
Only 6% of executives confirm clear, organization-wide AI ROI despite 89% reporting speed gains. Corroborates BCG’s 5% substantial-gains finding and McKinsey’s 6% high-performer share from an independent measurement angle.
The 14% who cracked measurable team-level ROI are characterized by three behaviors: AI use upstream in planning and prioritization (5.6x more likely), AI for collaboration improvement (9.4x more likely), and high worker trust in AI for information surfacing (2.3x more likely). Consistent with Stage 3+ maturity markers in BCG/MIT CISR models.
55% of executives report AI is widening performance gaps between teams — Stage 2 organizations with workflow integration capture gains; Stage 1 teams generate more output into unchanged bottlenecks.

Source: research/07-adoption-challenges/atlassian-state-of-teams-2026.md · Jan–Feb 2026 · MEDIUM-HIGH · TIER 1

Gartner Data Foundations Investment Gap (n=353, Nov–Dec 2025)

Gartner’s survey of 353 data and analytics leaders provides an investment-differential framing for maturity: organizations with successful AI initiatives invest up to 4x more as a share of revenue in data quality, governance, and AI-ready talent than those reporting poor outcomes. TIER 1.

Only 39% of technology leaders are confident their current AI investments will deliver positive financial impact. The majority are spending without conviction — a Stage 1 characteristic.
Organizations at the highest data-capability maturity report up to 65% better business outcomes (revenue growth + cost optimization) than lower-maturity peers (self-reported, not audited financials).
The investment gap is structural, not motivational. Companies reporting poor AI outcomes are not failing for lack of ambition or model access — they are under-investing in data plumbing. Corroborated by Davenport/Return on AI Institute (n=1,006, March 2026): data quality produces a 2x ROI multiplier. Cloudera/HBR (n=230, March 2026): only 7% of enterprises describe their data as completely AI-ready.

Source: research/05-analyst-firms/gartner-data-foundations-ai-success-2026.md · Gartner, n=353 D&A and AI leaders, Nov–Dec 2025 · MEDIUM-HIGH · TIER 1

McKinsey State of Organizations 2026 — Maturity Is Rare: Only 1% of US C-Suite Report Mature Rollouts (n=10,018)

McKinsey’s second annual organizational survey (n=10,018, 15 countries, 16 industries, March 2026) provides the largest-sample maturity calibration in the corpus. TIER 1.

Only 1% of US C-suite leaders describe their AI rollouts as mature. This is the most direct empirical anchor for Stage 1 dominance in the US enterprise: 88% are deploying, 81% see no bottom-line impact, and almost none of the deploying organizations have progressed to operational maturity.
The leadership ownership gap is the primary maturity predictor: Only 14% of organizations have leaders who consistently champion AI with clear strategy; 1 in 6 has no C-suite AI owner. Maturity cannot advance without an accountable owner — every maturity model (MIT CISR, BCG, McKinsey QuantumBlack) identifies C-suite sponsorship as a Stage 2→3 prerequisite.
The $5:$1 investment ratio operationalizes Stage 3 requirements: Organizations applying $5 in people investment per $1 in technology are 4.3x more likely to sustain top-tier performance. This maps directly to the MIT CISR Stage 3–4 financial performance spread (+11.3pp to +17.1pp) — both point to organizational capability investment as the maturity-unlocking variable.
Only 30% reallocate resources enterprise-wide when priorities shift — a structural blocker for Stage 3 progression, which requires redirecting headcount and operating budgets to AI-transformed workflows rather than simply layering AI onto unchanged org structures.

Source: research/01-ai-native-landscape/mckinsey-state-of-organizations-2026.md · McKinsey State of Organizations, n=10,018, March 2026 · MEDIUM-HIGH · TIER 1

Evanta / Gartner C-Suite Leadership Perspectives 2026 (n=2,505, March 2026)

Three parallel Gartner C-level Communities surveys — CIO (n=990), CHRO (n=430), CISO (n=1,085) — provide the largest concurrent cross-functional snapshot of AI maturity priorities from senior decision-makers. TIER 1.

Only 1% of US C-suite describe AI rollouts as mature (McKinsey corroboration cited within the Evanta dataset) — while 66% of CIOs are investing in AI/ML and operationalization is the #2 CIO priority. The gap between investment and maturity is the defining characteristic of Stage 2: tools acquired, outcomes not yet embedded.
“Operationalizing AI” displaced “experimenting with AI” as the language of CIO priority — a terminology shift that maps directly to Stage 2→3 transition pressure in the adoption cycle. Organizations at Stage 3 are closing the language gap; those still running pilots are behind the CIO-consensus vocabulary.
CISO adoption of “Enable & Protect AI” as a newly introduced #1 priority reflects that security governance is now the Stage 2→3 gating criterion — not technology or budget. Organizations cannot standardize AI (Stage 3) without a security-approved deployment framework.
CHRO “Change Management & Workforce Resiliency” rising from #5 to #3 calibrates where most organizations sit: deployment is underway but workforce adaptation is lagging — the behavioral marker of Stage 2/early Stage 3 in the maturity curve.

Source: research/05-analyst-firms/evanta-gartner-clevel-leadership-perspectives-2026.md · Evanta/Gartner, n=2,505 (CIO/CHRO/CISO), March 2026 · MEDIUM-HIGH · TIER 1

Sandbox to Production: The 8-Month Pilot-to-Production Gap (Digital Applied n=650, Mar 2026)

Enterprise survey of 650 VP-level technology leaders (Feb–Mar 2026) quantifies the time cost of the Stage 2→3 transition — the hardest gap to close in the maturity curve. TIER 2.

Only 14% of organizations with active AI agent pilots have reached production scale — while 78% have pilots underway. Stage 2 is now the modal enterprise state.
Median timeline: 8 months from prototype to production access for business users. Of the 64% that attempted to expand beyond pilot, 72% have been stalled for 6 months or longer — a quantified definition of “stuck at Stage 2.”
Five root causes of scaling failure account for 89% of stalled transitions: integration complexity (63%), output quality degradation at volume (58%), missing monitoring (54%), unclear ownership (49%), insufficient domain training data (41%). None are model-quality problems — all are organizational and infrastructural.
Sector gap: financial services reaches production at 21%, healthcare at 8% — Stage 3 arrival is gated by regulatory density, not technology readiness. Executives in regulated sectors need to factor regulatory gate cycles into their maturity timelines explicitly.

Source: research/16-procurement-contracting/sandbox-to-production-time.md · Digital Applied, n=650, February–March 2026 · MEDIUM-HIGH · TIER 2

Gartner 2026 Hype Cycle for Agentic AI — The Peak as a Maturity Marker (April 2026)

Source: research/12-agent-workers/gartner-hype-cycle-agentic-ai-2026.md · Gartner, April 2026 · MEDIUM-HIGH · TIER 1

Gartner’s Hype Cycle positioning of agentic AI adds a market-maturity dimension to the organizational maturity data: where the technology sits relative to enterprise readiness determines how difficult the Stage 2→3 transition is for organizations chasing the cycle.

Agentic AI platforms sit at the Peak of Inflated Expectations with a 2–5 year timeline to mainstream adoption. Organizations deploying now are building on pre-mainstream infrastructure — higher governance burden, faster-moving platform risk, and less available talent.
Generative AI moved to the Trough of Disillusionment in 2026. Organizations that have not yet extracted measurable value from GenAI are now being pushed to add agentic capabilities on top of unresolved GenAI governance problems.
17% deployed → 81% planning within 2 years is the sharpest adoption-intention gap Gartner has measured for any emerging technology — and it is occurring before the governance solutions (agentic AI governance, agentic AI security, FinOps for agentic AI) have reached mainstream maturity.
>40% cancellation rate forecast by 2027. For maturity-model purposes, this is the agentic-specific expression of the Stage 2→3 failure rate: organizations attempting to move from pilot to production without the organizational infrastructure to sustain it.

The Hype Cycle’s governance profiles — agentic AI governance and agentic AI security both appearing as distinct, pre-Peak tracks — confirm that the maturity gap is a sector-wide constraint, not a firm-specific failure. The organizations that will reach Stage 3 on agentic AI faster are those treating governance build-out as the primary maturity investment, ahead of platform decisions.

Mayfield 6th Annual CXO Survey — The Governance Debt as Maturity Constraint (n=266, Fortune 50–Global 2000, January 2026)

Source: research/12-agent-workers/mayfield-agentic-enterprise-cxo-survey-2026.md · Mayfield, n=266 CXOs Fortune 50–Global 2000, January 2026 · MEDIUM-HIGH / TIER 1

The Mayfield CXO survey provides the clearest evidence of how governance debt functions as a maturity ceiling:

72% in production or pilots; 60% lack formal governance frameworks. Maturity stage and governance stage are decoupled — organizations are advancing deployment faster than the oversight structures that would allow them to sustain it.
Data readiness has been the #1 blocker for five consecutive years (58% of CXOs). The consistent finding across 6 annual surveys is that maturity is not primarily gated by technology — it is gated by data infrastructure and organizational structure.
50%+ are actively reallocating spend from legacy vendors to AI-native solutions. This procurement shift signals a maturity inflection: organizations moving from experimentation to committed transformation rather than incremental optimization.
91% plan to increase agentic AI budgets in 2026. Investment intent is broad; governance build-out is narrow. This gap is the primary constraint on maturity advancement at the Fortune 50–Global 2000 tier.

PwC 2026 Digital Trends in Operations — The Four-Variable Maturity Test (n=767, Jan–Feb 2026)

Source: research/07-adoption-challenges/pwc-digital-trends-operations-2026.md · PwC, n=767 US operations/supply chain leaders revenue ≥$100M, Jan–Feb 2026 · MEDIUM / TIER 1

PwC’s 2026 operations survey offers the sharpest four-variable maturity test in the corpus. An organization is mature only when it simultaneously achieves all four:

AI fully embedded enterprise-wide
No significant barriers to scaling autonomous agents
A collaborative and horizontal operating structure
Technology investments fully delivering expected results

Only 4% of 767 respondents clear all four bars. Breakdown of individual constraints: 27% have embedded AI enterprise-wide; only 37% are comfortable with autonomous agent deployment; only 41% operate with horizontal structure; 89% say tech investments haven’t fully delivered.

This compound-condition framing is the most actionable maturity diagnostic in the corpus for COO/CIO audiences: it converts the abstract question “how mature are we?” into four testable, independently addressable gaps. The 4% figure, while vendor-defined, is consistent with Cisco’s Pacesetter finding (13% globally meeting their own multi-variable maturity criteria) and Accenture UK’s 1-in-10 scaling finding — all pointing to a top decile of operationally mature organizations separating from the field.

AI Daily Brief — Super Intelligent Maturity Maps: Six-Dimension Q2 2026 Benchmark (480+ Studies, 150,000+ Respondents)

Source: research/13-multimodal-sources/ai-daily-brief/2026-04-xx-ai-maturity-maps-q2-enterprise-readiness-benchmarks.md · Super Intelligent / Nathaniel Whittemore, 480+ studies synthesized, 150,000+ survey respondents, 50+ countries, Q2 2026 · MEDIUM / TIER 1

The most granular function-level AI maturity diagnostic in the corpus. Six dimensions scored 1–5 across 10 enterprise functions (customer service, engineering, IT, sales, marketing, HR, operations, finance, legal, product).

The adoption embedding gap is universal. Every function-specific survey shows the same pattern: high claimed adoption, low depth. Sales is the starkest — 88% claim AI use, 24% have it in revenue workflows.
People are the bottleneck getting the least investment. 7 of 10 functions score “significantly behind” on the people dimension. Deloitte: 93% of AI spend on infrastructure, 7% on people.
Data is the floor constraint. 8 of 10 functions score 1 or 1.5 out of 5 on data readiness. Without proprietary context, organizations cannot move past basic assistive usage.
Finance is the governance exception. 69% of CFOs have advanced AI governance frameworks — decades of SOX compliance transferred directly. They govern well; they deploy poorly.
Only 3 functions reach “on track” on any dimension: customer service (deployment depth, systems), engineering (deployment depth, systems, people), IT (deployment depth, systems, people). The structural advantage: technical practitioners, measurable workflows.

McKinsey State of AI Trust 2026 — Governance-Specific Maturity Scale (n=~500, March 2026)

A dedicated RAI (Responsible AI) maturity assessment across five dimensions — the only survey in the corpus focused specifically on governance and trust maturity rather than deployment or productivity:

Average RAI score: 2.3 / 4 (2026), up from 2.0 (2025) — positive trajectory, bottom half of scale.
Only 33% at level 3+ — the level at which policies actually produce oversight rather than just documentation.
Five dimensions: Strategy | Risk management | Data and technology | Governance | Agentic AI governance (new in 2026).
Pattern: Data and technology dimensions advance fastest. Strategy, governance, and agentic AI governance lag. The gap between technical and organizational maturity is the defining finding.
Agentic AI governance is universally the weakest dimension — only ~30% mature. Consistent across all regions and industries.
$25M+ RAI investment threshold: organizations above it are far more likely to achieve >5% EBIT impact. Maturity is not just risk reduction — it correlates with financial performance.

Cross-reference: MIT CISR 4-stage maturity (Stage 3–4 = +11.3pp to +17.1pp financial performance). BCG: only 5% of organizations getting substantial financial gains. McKinsey RAI survey provides the governance-specific lens on why the 95% are not there yet.

Source: research/04-consulting-firms/mckinsey-state-of-ai-trust-2026.md — MEDIUM-HIGH / TIER 1