Research · Frameworks
What the AI transformation frameworks say, and what backs them.
Fifteen named frameworks and studies from consultancies, business schools, cloud vendors and standards bodies. For each: what it says, how good the evidence is, and where it stops.
The short version
The frameworks barely differ in what they say. They differ in how good their evidence is, and none has independent proof that adopting it causes P&L results. The best-supported pattern is a correlation: firms that redesign whole workflows, put a named owner and KPIs on every use case, and put most of the effort into people and process report much higher returns. Those firms are a small minority (about 6% in McKinsey's 2026 survey).
Six elements recur: a CEO-owned ambition tied to the core P&L; redesigning workflows instead of inserting tools; tracking value rigorously; AI-ready data and reusable platforms; people and change (BCG's 70%); and governance in proportion to risk.
For founder-led firms the evidence points to a light, value-first operating model on two or three core workflows, not a purchased framework. No framework has been validated for small firms or for Vietnam, and no AI executive programme has independently measured P&L outcomes. The only randomised trial with small firms (Kenya) found no average effect on revenue or profit from an AI mentor: high performers gained and low performers lost.
| Framework | What it says | Evidence | Limits | Date | Source |
|---|---|---|---|---|---|
| State of AI / AI high-performer practicesMcKinsey & Company (QuantumBlack) · ConsultancyThemes: Winners redesign the work; they do not just add a tool; The pilot-to-production funnel narrows hard; Small and founder-led firms adopt later and see less | Practices across strategy, talent, operating model, technology, data and scaling. High performers pursue growth as well as efficiency, redesign workflows, show visible leadership ownership and define human-in-the-loop processes. | Self-published. Annual survey. 37% of respondents attribute any EBIT impact to AI and high performers stay flat at about 6%. Large firms (1bn+ revenue) report scaling AI across the enterprise at 54% against about a third of smaller ones. | Self-reported attribution; heavy investment is a marker of high performance, not proof of cause. The 6% has not moved in two years, so the practices are not spreading quickly. | 25 Aug 2026 | Source ↗Page opened |
| Future-Built and the 10-20-70 ruleBoston Consulting Group · ConsultancyThemes: Winners redesign the work; they do not just add a tool; The pilot-to-production funnel narrows hard; Self-reported benefit outruns measured P&L impact | Maturity groups (future-built, scalers, laggards). Five strategies: CEO-led multiyear ambition, AI-first operating model, fit-for-purpose tech and data, workforce involvement and upskilling, rigorous value tracking. Effort split 10% algorithms, 20% technology and data, 70% people and process. | Self-published. 1,250 senior executives. Future-built firms show 1.7x revenue growth, 3.6x three-year shareholder return and 1.6x EBIT margin versus laggards. The 2026 index reports 48.5% of firms creating measurable value (secondary report). | Survey-defined maturity correlates with market outcomes but causality can run the other way. BCG sells the transformation it describes. The 2026 jump partly reflects a reframed segmentation. | 30 Sept 2025 | Source ↗Page opened |
| State of AI in the Enterprise (2026 edition)Deloitte AI Institute · ConsultancyThemes: Skills and training are the binding constraint; Agents can do more than firms have deployed; Winners redesign the work; they do not just add a tool | Progression from surface use to strategic reinvention, covering agentic AI, governance and workforce redesign. | Self-published. 3,235 leaders in 24 countries (fielded Aug-Sep 2025). Secondary summaries report 84% have not redesigned jobs around AI and 21% have mature agent governance (not checked against the PDF). | Sample skewed to firms on the leading edge of AI (Deloitte's own note). High ROI-attainment figures conflict with other studies mainly because definitions differ. | 2026 | Source ↗Confirmed from search results |
| AI Maturity Model (five levels) and AI-ready dataGartner · ConsultancyThemes: The pilot-to-production funnel narrows hard; Self-reported benefit outruns measured P&L impact; Agents can do more than firms have deployed | Five levels from planning to leadership. Differentiators: use-case selection on value and feasibility, governance, engineering practice, dedicated AI leaders, centralised strategy and data, multiple value metrics. | Self-published. 432 organisations. 45% of high-maturity firms keep AI projects running 3+ years against 20% of low-maturity firms. 63% of high-maturity firms run financial or ROI analyses. | Outcome is project longevity, not profit. Full model is paywalled. "60% of AI projects abandoned" is a forecast tied to AI-ready data, not an observed rate. | 30 Jun 2025 | Source ↗Page opened |
| Accountable Acceleration (annual enterprise adoption study)Wharton Human-AI Research with GBK Collective · Business schoolThemes: Self-reported benefit outruns measured P&L impact; Small and founder-led firms adopt later and see less | Three-stage arc: exploration (2023), experimentation (2024), accountable acceleration (2025). Levers are people, process and ROI measurement. | Self-reported survey. About 800 US enterprise decision-makers. 82% use GenAI weekly, 72% formally measure ROI, about three quarters report positive returns. Mid-sized and smaller firms report faster ROI. | A benchmark, not a framework tested for outcomes. "Positive ROI" is respondent-judged and often productivity-based. | 28 Oct 2025 | Source ↗Page opened |
| Enterprise AI Maturity ModelMIT Center for Information Systems Research · AcademicThemes: The pilot-to-production funnel narrows hard; Winners redesign the work; they do not just add a tool | Four stages: experiment and prepare; build pilots and capabilities; develop AI ways of working; become AI future-ready. Moving from stage 2 to 3 needs strategic alignment, modular platforms and data, AI-ready people and teams, and stewardship. | Independent (corporate-funded). Survey of 721 firms plus 16 executive interviews. Stages 1-2 below industry-average financial performance, stages 3-4 well above. The 2025 update says the biggest gain comes from moving from stage 2 to 3, which is the pilot-to-P&L step. | Cross-sectional and correlational. Survey predates generative AI (2022). Large-enterprise sample. | 19 Dec 2024 | Source ↗Page opened |
| GenAI Divide: State of AI in Business 2025MIT Media Lab Project NANDA · AcademicThemes: The pilot-to-production funnel narrows hard | Diagnostic rather than a framework. Pilots fail because tools do not learn or adapt to workflows; partnering with learning-capable tools reaches deployment more often than building internally; budgets skew to visible sales and marketing pilots. | Preliminary. Review of 300+ public initiatives, interviews with 52 organisations and 153 leader surveys. Headline: 95% of organisations see no measurable P&L return. | Small, non-random sample; narrow P&L definition; methodology described inconsistently. Treat the 95% as a signal, not a base rate. | 1 Jul 2025 | Source ↗Confirmed from search results |
| Root Causes of Failure for AI ProjectsRAND Corporation · Independent researchThemes: The pilot-to-production funnel narrows hard | Five anti-patterns: misunderstood problem, missing data, chasing technology over the problem, inadequate infrastructure, applying AI to problems too hard for it. | Independent. 65 experienced practitioners interviewed. 84% cited leadership-driven root causes as primary. | Practitioner perceptions only. The "80% fail" figure is quoted from outside work. Pre-agent era. | 13 Aug 2024 | Source ↗Page opened |
| Cloud Adoption Framework for AI and agentsMicrosoft · Cloud vendorThemes: Governance and liability catch up with deployment | Six phases: strategy, plan, ready, govern, secure, manage. Separate startup and enterprise checklists. Governance aligned to the NIST AI RMF. Agent adoption guidance added Dec 2025. | Vendor. No independent outcome evidence found. | Platform-centric. Strong on technical readiness, weak on P&L value capture. The startup checklist is one of few size-tiered tracks in any framework. | Living document | Source ↗Confirmed from search results |
| Cloud Adoption Framework for AI, ML and generative AIAmazon Web Services · Cloud vendorThemes: Governance and liability catch up with deployment | Six perspectives: business, people, governance, platform, security, operations, with AI-specific capabilities ordered by maturity and a responsible-use capability. | Vendor. No independent outcome evidence found. | A capability checklist rather than a value-realisation method; same platform bias as other cloud frameworks. Current version date not confirmed. | Living document | Source ↗Confirmed from search results |
| AI Risk Management Framework 1.0US National Institute of Standards and Technology · Governance standardThemes: Governance and liability catch up with deployment | Four functions: govern, map, measure, manage, plus a playbook and a generative AI profile (2024). Voluntary. | Widely referenced. No study found linking adoption to financial results. | A risk framework, not a value framework. Useful to size controls in proportion to risk. | 1 Jan 2023 | Source ↗Page opened |
| ISO/IEC 42001 AI management systemISO/IEC JTC 1/SC 42 · Governance standardThemes: Governance and liability catch up with deployment | Certifiable plan-do-check-act management system: policy, roles, risk and impact assessment, lifecycle, data, transparency and third-party controls. | Certification exists. First major cloud-provider accredited certification reported Nov 2024. A 2026 arXiv review notes adoption evidence remains thin. | Risk of governance theatre. Value is mainly customer and procurement trust, not P&L. Worth considering only if enterprise customers require it. | 1 Dec 2023 | Source ↗Confirmed from search results |
| AI Governance Alliance playbooksWorld Economic Forum · InstitutionThemes: Governance and liability catch up with deployment; The pilot-to-production funnel narrows hard | Nine plays for responsible AI innovation (Sep 2025); a four-action framework for transformational AI in financial services (Jun 2026): board-accountable vision, systems, near-term vs long-term balance, workforce and culture. | Case-based. Built from showcase cases (financial-services playbook draws on 150+ executives from 100+ organisations). | Consensus documents with selection bias. Checklists, not evidence. | 22 Sept 2025 | Source ↗Confirmed from search results |
| AI adoption by small and medium-sized enterprisesOECD · InstitutionThemes: Small and founder-led firms adopt later and see less; Skills and training are the binding constraint | Segments SMEs into adoption stages with stage-specific barriers and policy supports. | Independent survey. 5,000+ SMEs in seven countries. Skills are the top barrier and cost is not. Only 23.6% of GenAI-using SMEs train staff on AI. | Policy-oriented, not a firm-level playbook. No P&L outcome data. | 1 Dec 2025 | Source ↗Confirmed from search results |
| ADKAR applied to AI adoptionProsci · Change managementThemes: Skills and training are the binding constraint | Individual change through awareness, desire, knowledge, ability and reinforcement, plus an AI adoption diagnostic. | Vendor. 1,107 professionals surveyed: 63% cite human factors as a primary challenge; user proficiency is the top issue (38%). | Diagnoses adoption, not value. Prosci sells the method and certification. Publication date not confirmed. | Living document | Source ↗Confirmed from search results |
Widely quoted failure rates are weaker than their circulation suggests. MIT NANDA's "95%" rests on a small preliminary sample; RAND's "80%" is quoted from outside work, not measured by RAND; Gartner's "60% abandoned" is a forecast. Use them as directional signals. Not assessed in this pass: Google Cloud, Kotter, Bain, INSEAD, IMD, Stanford HAI and HBR frameworks. Vietnam data on adoption come mostly from vendor-sponsored surveys with inconsistent definitions. See the Knowledge graph for how each framework links to the evidence themes.