Integrated Assessment

Most AI programs are funded one track at a time — whichever function shouted loudest — and then stall for reasons that live in a track nobody scored. The Integrated Assessment exists to stop that. It is a single cross-functional working session that scores all eight tracks on one shared ladder, lays the results side by side, and produces a prioritized gap list: not "here is everything wrong," but "here is the one constraint to relieve next." It is the operational front door to Running the Program — you run it before you sequence, and again on a cadence to see whether the constraint has moved.

Run it interactively. The AI Maturity Self-Assessment is a self-scoring version of this diagnostic: score each track's underlying dimensions, watch the eight-track profile build on a live radar, and get the binding-constraint and profile-shape read (spiky · flat-but-low · foundation-gap) computed for you. Use it to prepare for — or stand in for — the working session described below.

This page is the how-to-run-it companion to the Framework Architecture, which defines the eight tracks, their dependencies, and the two failure modes, and to the Maturity Model, which defines the shared five-level ladder and the full maturity matrix. This page does not re-print that matrix; it tells you how to source a score for each track, read the pattern, and turn it into an investment case. The need is real: only 13% of organizations are "fully ready" to deploy and leverage AI — down from 14% the year before (Cisco, 2024), only 1% of leaders describe their AI deployment as "mature" (McKinsey, 2025), and at least 30% of gen-AI projects are projected to be abandoned after proof of concept by the end of 2025 — for reasons Gartner lists as poor data quality, inadequate risk controls, escalating costs, and unclear business value (Gartner, 2024). Every one of those failure modes lives in a different track. A single-track view cannot see them; an integrated one can.

What This Is For¶

The Integrated Assessment answers one question: given everything in motion, where is value actually being throttled, and what should we fund next? It is deliberately a cross-functional session, not a survey emailed to eight owners, because the most expensive AI gaps are organizational, not technical. BCG's guidance is to direct roughly 70% of AI effort and resources to people and process, 20% to technology and data, and 10% to algorithms (BCG, 2024) — and you only find people-and-process gaps when the people who own them are in the same room reconciling their scores against each other.

Use it three ways:

At program start — to set the baseline and the first sequence (feeds Sequencing Playbooks and The 90-Day Launch).
On a cadence — quarterly or per planning cycle, to confirm the binding constraint has moved and to re-prioritize. Organizations that keep AI on a recurring governance rhythm pull ahead: 63% of high-AI-ROI organizations discuss AI at every board meeting, versus only 13% of low-ROI organizations (Protiviti, 2026).
Before a major investment decision — to check that the spend targets the actual constraint, not the most visible layer.

The Eight Dimensions Are the Eight Tracks¶

The diagnostic has exactly eight dimensions because the framework has exactly eight tracks. There is no separate instrument to maintain: each track owns a detailed assessment, and the Integrated Assessment is the roll-up that places those eight scores on one comparable scale. In the session you do not re-derive each score from scratch — you pull it from the track's own assessment page, or run that assessment if it has never been scored.

#	Dimension (Track)	What you are scoring	Source assessment
01	AI Strategy & Leadership	Is there a funded, business-anchored set of AI priorities every other track is scoped against?	Strategic Readiness Scoring
02	AI Governance & Risk	Do guardrails — acceptable use, model risk, compliance — exist and gate real work?	Governance Maturity Scoring
03	Data Readiness	Is the data AI depends on trustworthy, governed, and accessible to use cases?	AI Readiness Assessment Framework
04	Technology Architecture & Platform	Is there a coherent platform with standard model access and patterns, not point-solution sprawl?	Platform Maturity Scoring
05	Workflow Optimization & Automation	Are end-to-end workflows actually being redesigned with AI, not just personal assistance?	Workflow Maturity & Opportunity Scoring
06	AI Adoption & Culture	Do target teams use AI in daily work, with real behavior change rather than shelfware?	Adoption Maturity Scoring
07	Talent & Capability Building	Does the organization have — and keep building — the competencies the work requires?	Talent Readiness Scoring
08	Measurement & Value Realization	Can you tie AI to value, prove it to finance, and let evidence reprioritize spend?	Measurement Maturity Scoring

The eight group into four layers — Direction (AI Strategy & Leadership, AI Governance & Risk), Foundation (Data Readiness, Technology Architecture & Platform), Execution (Workflow Optimization & Automation, AI Adoption & Culture, Talent & Capability Building), and Feedback (Measurement & Value Realization) — which is the lens you use later to read the pattern. Each track's assessment uses its own dimensions internally (Measurement, for example, scores five sub-dimensions); the Integrated Assessment consumes only the headline level each produces.

Scoring Methodology¶

Every track is scored on the same five-level ladder defined in the Maturity Model, so the eight scores are directly comparable:

1 Nascent · 2 Developing · 3 Emerging · 4 Scaling · 5 Transformational

Two rules make the scores honest:

Score on evidence, not intent. The level is what is demonstrably true today — backed by an artifact a skeptic would accept (a funded plan, a deployed control, a production workflow, a finance-reviewed value number) — never what is planned, budgeted, or "almost done." A roadmap to redesign workflows is not a redesigned workflow. This is the single most common way an assessment lies to its sponsor.
Source each score from the track's own assessment, then reconcile in the room. Each owner brings their track's score with its evidence. The session's job is to challenge it: a Strategy owner claiming Level 4 while no other track can name the three funded priorities is really at Level 2. The cross-functional table is the calibration mechanism — it catches the optimism that a self-scored survey never would.

Record one level per track plus a one-line evidence note. The output is a profile of eight numbers, not an average — averaging is the cardinal error, because it hides exactly the low score that is throttling everything else.

Reading the Pattern: Finding the Binding Constraint¶

A completed profile is read for shape, not sum. The whole point of scoring all eight at once is to see which one is gating the rest.

The binding-constraint principle: a track is only as effective as its weakest upstream dependency. Funding Workflow Optimization & Automation on top of unready Data Readiness produces motion without value; pushing AI Adoption & Culture without Talent & Capability Building produces logins without usage. Find the lowest track that is throttling the ones downstream of it — and fund that, not the most visible or fashionable layer.

Three patterns to look for, drawn from the Maturity Model:

Spiky (a 4 next to a 1). A high track is being throttled by a low neighbor. This is the binding constraint made visible. The 03 → 04 (Data → Platform) and 05 → 06/07 (Workflow → Adoption/Talent) boundaries are where programs most often stall — "working pilots that won't scale" almost always trace to a Level-1 dependency next door. Each of the three classic constraints has its own evidence: data quality and availability is the top-cited AI adoption barrier at 52% (PEX Network, 2025); 63% of employers name skills gaps the biggest barrier to transformation through 2030 (WEF, 2025); and regulatory compliance is the #1 barrier to deploying gen-AI, cited by 38%, up from 28% a year earlier (Deloitte, 2025). Whichever is your lowest score is your constraint.
Flat-but-low (everything at 2). No track has crossed the threshold where value appears — the modal organization. The fix is not to nudge all eight up a notch. Pick one or two tracks on the critical path and push them to Level 3–4 so something crosses into value, then re-assess.
Foundation gap under execution pressure. High Direction and Execution scores sitting on Level-1/2 Foundation (Data, Platform) is the "boil the ocean" failure pre-loaded — execution work is being built on substrate that will force a redo. Relieve Foundation before pouring more into Execution.

The constraint you name here is the input to Sequencing Playbooks: sequence by dependency, not by track number.

Framing the Investment Case¶

A prioritized gap list is only useful if it changes where money goes. Convert the profile into a case in three moves, anchored to Measurement & Value Realization:

Name the constraint and its downstream cost. "Data Readiness at Level 2 is capping three Level-4 workflow pilots that cannot scale." That sentence — constraint, level, and the higher-scoring work it throttles — is the case. It reframes the ask from "fund Data" to "unlock the value already half-built on top of it."
Size the prize against the leader–laggard gap. The reason to relieve the constraint is the value differential it unlocks. BCG's "future-built" leaders — just 5% of organizations — report roughly 1.7x the revenue growth, 1.6x higher EBIT margins, and 40% greater cost savings than laggards (BCG, 2025; BCG, 2025), and only about 6% of organizations are AI high performers attributing more than 5% of EBIT to AI (McKinsey, 2025). The gap between the profile you have and the one those firms have is the size of the opportunity.
Commit to measuring the unlock. Tie the investment to a metric from Measurement & Value Realization and a re-assessment date, so the next session can show the constraint moved. This is what separates leaders: more than a third of gen-AI high performers have clear ways to measure and track the value of their gen-AI work, versus only about one in ten of other organizations (McKinsey, 2024). Without the measurement commitment, the gap list becomes a wish list.

Run It in 90 Minutes¶

One session, one room (or one call), all eight track owners plus the program sponsor. Pre-work: each owner brings their track's current score with evidence.

Time	Activity
0–10 min	Frame. Sponsor restates the question — where is value being throttled, and what do we fund next? Reaffirm: score on evidence, not intent.
10–45 min	Score the eight tracks. Each owner presents their level and its evidence (~4 min each). The room challenges and calibrates; record one level + a one-line note per track.
45–60 min	Plot the profile. Lay the eight numbers in track order. Mark the layers (Direction / Foundation / Execution / Feedback). Look for spiky, flat-but-low, and foundation-gap patterns.
60–75 min	Name the binding constraint. Agree the single lowest track throttling downstream work. Resist the urge to fix everything — name one (at most two).
75–90 min	Frame the case + commit. State constraint → downstream cost → the unlock. Assign an owner, a first move (hand off to Sequencing Playbooks), a metric from Measurement & Value Realization, and the re-assessment date.

Keep it to 90 minutes by timeboxing ruthlessly: the deep work already happened inside each track's own assessment. This session reconciles and prioritizes — it does not re-derive.

Key Takeaways¶

One session, one ladder, eight tracks. The Integrated Assessment scores all eight tracks on the shared 1 Nascent · 2 Developing · 3 Emerging · 4 Scaling · 5 Transformational ladder and produces a prioritized gap list — not a to-do list.
The eight dimensions are the eight tracks, 1:1. Source each score from the track's own assessment page; the session reconciles and rolls up, it does not re-derive.
Score on evidence, not intent, and never average. A profile of eight numbers tells the truth; a mean hides the low score that is throttling everything. Only 1% of leaders call their AI deployment mature (McKinsey, 2025) — optimism is the default failure.
Read the shape to find the binding constraint. Spiky (4-next-to-1), flat-but-low, and foundation-gap patterns each point to the one track to fund next — usually data, talent, or governance, the three most-cited blockers (PEX Network, 2025; WEF, 2025; Deloitte, 2025).
Convert the constraint into an investment case and commit to measuring the unlock. Tie it to a metric from Measurement & Value Realization and a re-assessment date — high performers are several times more likely than others to have clear ways to measure gen-AI value (McKinsey, 2024), and they run AI on a recurring cadence (Protiviti, 2026).

Sources¶

Cisco — Cisco 2024 AI Readiness Index, 2024. Only 13% of companies today are fully ready to capture AI's potential — down from 14% a year ago. View source · verified 2026-06-20 · primary
McKinsey — Superagency in the Workplace, 2025. Only 1 percent of leaders call their companies mature on the deployment spectrum, meaning AI is fully integrated into workflows and drives substantial business outcomes. View source · verified 2026-06-20 · primary
Gartner — Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept By End of 2025, 2024. At least 30% of generative AI projects will be abandoned after proof of concept by the end of 2025, due to poor data quality, inadequate risk controls, escalating costs or unclear business value. View source · verified 2026-06-20 · primary
BCG — Where's the Value in AI?, 2024. AI leaders follow the rule of putting 10% of their resources into algorithms, 20% into technology and data, and 70% into people and processes. View source · verified 2026-06-20 · primary
Protiviti — 2026 Global Board Governance Survey (with BoardProspects), 2026. In 63% of organizations reporting high AI ROI, every board meeting agenda includes a discussion on AI, compared with only 13% of low-ROI organizations. View source · verified 2026-06-20 · primary
PEX Network — PEX Report 2025/26, 2025. More than half of respondents (52%) cite data quality and availability as the biggest AI adoption barrier. View source · verified 2026-06-20 · primary
World Economic Forum — Future of Jobs Report 2025, 2025. The skills gap continues to be the most significant barrier to business transformation today, with 63% of employers already citing it as the key barrier they face. View source · verified 2026-06-20 · primary
Deloitte — The State of Generative AI in the Enterprise: Generating a new future (Q4), 2025. Regulatory compliance emerged as the top barrier to developing and deploying GenAI, increasing from 28% (Wave 1) to 38% (Wave 4). View source · verified 2026-06-20 · primary
BCG — The Widening AI Value Gap, 2025. Only about 5% of companies (the 'future-built') are generating value at scale, while 60% of companies are laggards with little or no value; 35% are in between. View source · verified 2026-06-20 · primary
BCG — The Widening AI Value Gap, 2025. Future-built companies already generate 1.7 times more revenue growth and 1.6 times higher EBIT margins than laggards, with about 40% more cost savings. View source · verified 2026-06-20 · primary
McKinsey — The State of AI, 2025. Respondents who attribute EBIT impact of 5 percent or more to AI use and say their organization has seen significant value from AI — about 6 percent of respondents — are defined as AI high performers. View source · verified 2026-06-20 · primary
McKinsey — The State of AI in early 2024, 2024. More than a third of gen-AI high performers have clear ways to measure and track the value of their gen-AI work, versus only about one in ten of other organizations; this specific split could not be confirmed against a public McKinsey 2024 quote and should be re-verified. View source · verified 2026-06-20 · ⚠ primary source unreachable