Primary-Verifying the Big-4 Readiness Frameworks: One Claim Holds, One Is Wrong, and McKinsey's "AI Readiness Index" Does Not Exist

The question

"Primary-verify the Big-4 agentic framework specifics — McKinsey's AI Readiness Index dimensions, Accenture's AI Navigator, KPMG's 10 pillars — against primary pages rather than search summaries."

Context: [[2026-07-15-agentic-assessment-framework-competitive-landscape]] carried a Big-4 paragraph sourced from search summaries and one aggregator blog, self-tagged "do not quote to clients until primary-verified." This brief closes that tag, claim by claim.

What we already know (from the vault)

• The parent brief's Big-4 paragraph makes five discrete factual claims, all flagged secondary: (1) McKinsey Rewired is a six-capabilities framework; (2) McKinsey publishes an AI Readiness Index scoring strategy / data / technology / organization / capabilities; (3) Accenture AI Navigator generates maturity scores across data / talent / process / technology with partial automation; (4) KPMG publishes a formally numbered 10-pillar framework; (5) that KPMG framework carries an ISO/IEC 42001 claim. See [[2026-07-15-agentic-assessment-framework-competitive-landscape]].
• The single named source for claims 2, 3 and 4 was a third-party aggregator, consulting-huber.com, not any vendor page. That is the weakest possible provenance for a competitor claim.
• The parent brief built a load-bearing synthesis argument on top of those claims: that "the Big-4's five dimensions are pre-agentic" and score no autonomy axis, which is what made autonomy a differentiator worth elevating. That argument inherits whatever the claims' error bars are.
• A qmd sweep of the vault found no other doc discussing McKinsey's, Accenture's or KPMG's readiness frameworks. The parent brief is the sole carrier, so correcting it corrects the vault.
• phData sits deliberately a tier below the Big-4 at the $150-500k mid-market band, so these frameworks are reference points and citation material rather than head-to-head bid competition. See [[2026-05-21-enterprise-ai-agent-deployment-paths]].

What the web says

Verification retrieved 2026-09-26. Method note: mckinsey.com blocked both curl and WebFetch (connection reset / 60s timeout) and was reached with one Playwright attempt; accenture.com's insights page is JS-rendered and needed WebFetch; kpmg.com and newsroom.accenture.com served to plain curl with a browser user-agent.

• Claim 1, McKinsey Rewired six capabilities: CONFIRMED. McKinsey's own article states "capturing this kind of value from digital and AI requires building six critical enterprise capabilities," and its Exhibit 2 is captioned "Six enterprise capabilities are critical for successful digital and AI transformations." The six are nameable from the article's own section headings: a business-led road map aligned across the C-suite, talent bench, an operating model that can scale, technology for speed and distributed innovation, data embedded everywhere, and adoption and scaling. Caveat that matters: this is dated June 20 2023 and is framed as digital and AI transformation, not agentic readiness. (McKinsey, "Rewired to outcompete") (primary-verified)
• Claim 2, McKinsey "AI Readiness Index": NOT FOUND ON A PRIMARY SOURCE — and the absence is well-evidenced, not just an unsuccessful search. A site search on mckinsey.com for the quoted phrase "AI readiness index" returns 869 loosely-matched results and zero containing the exact phrase. The phrase also appears zero times on the Rewired page, which contains no instance of the word "index" at all. Neither the name nor the five dimensions attributed to it can be sourced to McKinsey. The nearest genuine assets are differently-shaped: "AI data readiness" (six data readiness disciplines for chief data officers, 2026-06-23) and "From adoption to impact: Three horizons of AI transformation," which discusses an "AI readiness gap" (2026-07-08). Most likely the aggregator fused Rewired with generic readiness vocabulary and coined an index McKinsey never published. (McKinsey site search) (primary-verified negative)
• Claim 3, Accenture AI Navigator as a maturity-scoring assessment: CONTRADICTED. AI Navigator for Enterprise is real, but Accenture describes it in its own wording as "a generative AI-based platform that will help clients define business cases, make decisions, navigate AI journeys, choose architectures and understand algorithms." The page carries no maturity-score language, no data/talent/process/technology dimension set, and no assessment-automation claim. It is decision-support and build tooling, not a diagnostic. (Accenture, AI Navigator for Enterprise) (primary-verified)
• Claim 3b, what Accenture's readiness framework actually is — and the supersession the parent brief missed. On 2026-06-08, a month before the parent brief was written, Accenture and the Carnegie Mellon University Software Engineering Institute launched the AI Adoption Maturity Model. Accenture's words: it "divides AI-relevant capability areas into eight core dimensions: organizational strategy, workforce and culture, workflow re-engineering, risk and governance, data, engineering, operations and ecosystem," organized around "two key, high-level goals: organizational change and AI lifecycle engineering," and shipping with "an assessment tool that enables structured implementation and benchmarking of client outcomes across industries." Manish Sharma, Chief Strategy and Services Officer, frames it as meeting organizations "across eight critical dimensions of AI readiness." (Accenture newsroom, 2026-06-08) (primary-verified)
• Claim 4, KPMG's 10 pillars: PARTIALLY CONFIRMED, with a category error underneath. The count is exactly right. KPMG's Trusted AI framework page enumerates ten: Reliability, Security, Safety, Privacy, Sustainability, Explainability, Data, Transparency, Fairness, Accountability. Two corrections though. They are not numbered on the page, they are an unnumbered list, so "formally numbered" overstates it. More importantly KPMG calls them "the KPMG Trusted AI ethical pillars" and the page contains zero instances of "readiness" and zero of "maturity." This is a responsible-AI governance framework, not an AI-readiness maturity model, so placing it in a readiness-framework lineup misfiles it. (KPMG Trusted AI framework) (primary-verified)
• Claim 5, the ISO/IEC 42001 claim attached to KPMG's 10 pillars: NOT FOUND ON A PRIMARY SOURCE. "42001" appears zero times on the Trusted AI framework page. KPMG member firms do publish ISO/IEC 42001 material (KPMG Switzerland, KPMG Canada) but as separate advisory and assurance offerings, with no primary text tying certification to the ten pillars. The claim as the parent brief phrased it stays unquotable.

Convergences and contradictions

• Scorecard: of five claims, one is CONFIRMED, one is PARTIAL, one is CONTRADICTED, and two are NOT FOUND on any primary source. Only the McKinsey Rewired six-capabilities claim survives intact. That is a 20% clean-pass rate on a paragraph that was already self-flagged, which retroactively justifies the flag and argues for treating aggregator-sourced competitor detail as unusable rather than provisional.
• The contradiction that costs the most: the parent brief's "the Big-4's five dimensions are pre-agentic" line is now materially weakened. Accenture's real framework has eight dimensions including workflow re-engineering, engineering, operations and ecosystem, sitting on CMU SEI's CMMI lineage. That is a lifecycle-engineering maturity model, not a pre-agentic executive quiz. It is still fair to say none of the three publicly scores per-agent authority or autonomy level, but the framing has to shift from "they are pre-agentic" to "they score engineering and organizational discipline, not per-agent authority." The wedge is narrower and needs different words.
• Convergence, and a mild validation signal: the Accenture/SEI model's two-goal spine, "organizational change" and "AI lifecycle engineering," is close to an independent restatement of the split phData's Organizational Intelligence framing already makes between the Org Platform and Intelligence Platform halves. Two teams landing on the same seam is weak but real evidence the seam is the right one.
• No renames or retirements to report on the three frameworks themselves. Rewired is live and still promoted, though McKinsey Digital URLs now resolve under a renamed /capabilities/tech-and-ai/ taxonomy. KPMG Trusted AI is live and states "we will continue to review and evolve our approach." AI Navigator for Enterprise is live but was never the readiness asset. The one genuine change is additive: Accenture's AI Adoption Maturity Model, 2026-06-08.

Synthesis for RDCO

The operative output of this brief is a deletion, and it should be treated as the deliverable rather than as a disappointment. Three of the five Big-4 claims the vault was carrying cannot be primary-sourced, so they come off any phData Organizational Intelligence artifact that a client, a partner, or a phData account lead might read. The highest-risk one is specific enough to name: do not ever write or say "McKinsey's AI Readiness Index." It is not a McKinsey product. McKinsey's own site search cannot find the phrase. Competitive-landscape slides get read by buyers who used to work at these firms, and a confidently-cited framework that the alleged publisher has never heard of does not cost you one bullet, it costs the credibility of every other number on the page. Same disposal for the ISO/IEC 42001 attachment to KPMG's pillars, and for the "AI Navigator generates maturity scores" line, which is not merely unsourced but wrong about what the product is.

What replaces the deleted paragraph is shorter, stronger, and now citable. Three verified facts carry the same argumentative load: McKinsey's Rewired names six enterprise capabilities and dates to June 2023; Accenture and Carnegie Mellon SEI launched an eight-dimension AI Adoption Maturity Model on 2026-06-08; KPMG's Trusted AI framework is ten ethical pillars and contains no readiness or maturity language at all. That last correction is quietly the most useful of the three for the OI work, because it means the governance vocabulary and the readiness vocabulary in this market are two different conversations held by different frameworks. OI's governance surface can map itself against KPMG's ten named pillars as a coverage checklist, with a primary citation behind it, and that is a cheap credibility win that costs no original research.

The substantive strategic update is that Accenture's real framework is a closer competitor to Org Map and OIP than anything the July brief identified, and it should be studied rather than waved at. An eight-dimension model spanning organizational strategy, workforce and culture, workflow re-engineering, risk and governance, data, engineering, operations and ecosystem, explicitly split across "organizational change" and "AI lifecycle engineering," with a benchmarking assessment tool attached and CMU SEI's maturity-modeling authority behind it, is the most direct public analog to what OI is assembling. The right read is not threat, it is calibration. It confirms the Org Platform / Intelligence Platform division is a seam serious people converge on, and it sets the bar for what a defensible dimension set looks like. It also means any OI artifact claiming novelty for the shape of an org-plus-engineering readiness split is now claiming something a Fortune-500 buyer may have already seen from Accenture three months earlier.

Consequently the autonomy axis needs rewording, not abandoning, and this is the one place the parent brief's strategy still holds. None of the three verified frameworks scores how much authority an individual agent may hold. That gap is real and IMA's per-agent autonomy routing still sits in it. But the old sentence, that the Big-4 dimensions are pre-agentic five-dimension executive scorecards, is no longer accurate against Accenture and would be easy for an informed listener to knock down. The honest version is narrower and harder to rebut: these frameworks score organizational and engineering maturity at the enterprise level, and none of them scores authority at the agent level, which is exactly the unit an agentic estate has to govern. Framed that way the claim survives contact with the Accenture/SEI model instead of being embarrassed by it. Note for scope: RDCO is dormant as of 2026-09-23, so none of this is service-line positioning. It is input to phData OI artifacts and to how Ben talks about OI in phData rooms.

Why this is in the vault

This retires the explicit "do-not-quote-to-clients-until-primary-verified" tag on the Big-4 paragraph in [[2026-07-15-agentic-assessment-framework-competitive-landscape]], and it changes what goes into phData Organizational Intelligence client-facing material in two concrete ways: it removes a fabricated McKinsey product name that would have been cited in a competitive-landscape slide, and it forces the rewording of the autonomy-axis differentiation claim that the July brief recommended elevating to a client-visible scoring axis.

Open follow-ups

• Is the Accenture/SEI AI Adoption Maturity Model's practice-level rubric public or partner-gated? The eight dimensions are published, but maturity is "measured by how well key practices across these dimensions are implemented." If those practice statements are public, they are the sharpest available external benchmark for Org Map's scoring rubric. If gated, OI's rubric is unobserved by competitors too.
• Does any Big-4 framework score agent-level authority in material that is not on a landing page? Absence was verified on public pages only. The gated PDFs and assessment tools could contain an autonomy dimension, which would remove OI's clearest remaining differentiator.
• Deloitte's "executive's framework for AI capability assessment" is still secondary-only. It was outside this brief's three-vendor scope but sits in the same do-not-quote bucket in the parent brief, and it is the last unverified Big-4 claim the vault carries.

Related

• [[2026-07-15-agentic-assessment-framework-competitive-landscape]] — the parent brief this verifies and corrects; its Big-4 paragraph and its "pre-agentic dimensions" synthesis claim both change as a result
• [[2026-06-28-snowflake-si-cortex-positioning-caf-gap]] — the original competitor-positioning brief that started the assessment-framework thread
• [[2026-05-21-enterprise-ai-agent-deployment-paths]] — places phData a tier below the Big-4, which is why these frameworks are citation material rather than bid competition
• [[2026-05-20-phdata-cortex-agents-practice]] — the phData delivery spine that Organizational Intelligence artifacts feed
• [[2026-06-09-consultancy-partnership-tracks-velocity]] — reusable IP as the rewarded unit, the reason a defensible dimension set matters

Sources

Vault:

• [[2026-07-15-agentic-assessment-framework-competitive-landscape]] — ~/rdco-vault/06-reference/research/2026-07-15-agentic-assessment-framework-competitive-landscape.md
• [[2026-06-28-snowflake-si-cortex-positioning-caf-gap]] — ~/rdco-vault/06-reference/research/2026-06-28-snowflake-si-cortex-positioning-caf-gap.md
• [[2026-05-21-enterprise-ai-agent-deployment-paths]] — ~/rdco-vault/06-reference/research/2026-05-21-enterprise-ai-agent-deployment-paths.md
• [[2026-05-20-phdata-cortex-agents-practice]] — ~/rdco-vault/06-reference/research/2026-05-20-phdata-cortex-agents-practice.md
• [[2026-06-09-consultancy-partnership-tracks-velocity]] — ~/rdco-vault/06-reference/research/2026-06-09-consultancy-partnership-tracks-velocity.md

Web (primary-verified — vendor's own pages, retrieved 2026-09-26):

• McKinsey, "Rewired to outcompete," pub. 2023-06-20 — six critical enterprise capabilities; contains no "index" and no "AI Readiness Index": https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/rewired-to-outcompete (retrieved via Playwright; curl and WebFetch both blocked)
• McKinsey site search, exact phrase "AI readiness index" — 869 fuzzy results, zero exact-phrase matches: https://www.mckinsey.com/search?q=%22AI+readiness+index%22
• Accenture, AI Navigator for Enterprise announcement — "a generative AI-based platform that will help clients define business cases, make decisions, navigate AI journeys, choose architectures and understand algorithms"; no maturity-score language: https://www.accenture.com/us-en/insights/accenture-invest-3-billion-ai-accelerate-clients-reinvention
• Accenture newsroom, "Accenture and the Carnegie Mellon University Software Engineering Institute Launch AI Adoption Maturity Model," pub. 2026-06-08 — eight core dimensions, two high-level goals, attached assessment tool: https://newsroom.accenture.com/news/2026/accenture-and-the-carnegie-mellon-university-software-engineering-institute-launch-ai-adoption-maturity-model-to-help-organizations-scale-ai-with-predictable-outcomes
• KPMG, Trusted AI framework — ten ethical pillars enumerated (Reliability, Security, Safety, Privacy, Sustainability, Explainability, Data, Transparency, Fairness, Accountability); zero instances of "readiness," "maturity," or "42001": https://kpmg.com/xx/en/what-we-do/services/ai/trusted-ai-framework.html

Claims left unquotable (no primary source found; do NOT cite to clients):

• McKinsey "AI Readiness Index" and its alleged strategy / data / technology / organization / capabilities dimensions — no such named McKinsey asset
• ISO/IEC 42001 as a property of KPMG's ten Trusted AI pillars — absent from the framework page
• Accenture AI Navigator as a maturity-scoring assessment across data / talent / process / technology — contradicted by Accenture's own description
• Deloitte's "executive's framework for AI capability assessment" — out of scope here, still secondary-only in [[2026-07-15-agentic-assessment-framework-competitive-landscape]]