1. Decision Intent

What this question diagnoses: Whether leadership has moved from AI curiosity to AI accountability — i.e., whether there is a named owner for each AI outcome, a budget line, and a measurable gap it is closing. The 70% failure zone begins here: firms that cannot articulate the decision they are trying to improve cannot govern, measure, or scale it

Our leadership acknowledges AI is important, but no specific business problem has been formally scoped or assignedWe have AI ambitions documented in strategy decks, but no use case has a named owner, baseline metric, or funded mandatePriority use cases are defined with expected outcomes, and at least one has a cross-functional owner and a 12-month delivery targetAI goals are directly mapped to board-level P&L levers — each initiative has a named accountable executive, a baseline KPI, and a quarterly review cadenceAI investment decisions are governed like capital allocation: every initiative has a documented decision thesis, a cost-per-outcome target, and a kill/scale trigger

2. Data Supply Chain

What this question diagnoses: Whether data moves from source to decision-maker within the window where action is still possible — or whether it arrives too late to change the outcome. Most firms have data. Very few have decision-ready data at the moment of consequence

Data is owned by individual teams — there is no shared access model, and critical datasets require manual extraction requestsWe have a central data warehouse and dashboards, but data preparation for analysis still requires days to weeks of manual effortKey datasets are catalogued with ownership assigned; analysts can build analyses in hours rather than days using documented schemasAutomated pipelines deliver clean, versioned data with quality tracking and lineage documentation; data contracts exist between producers and consumersDecision-relevant signals reach the named decision-maker proactively — inside the operational window where the action can still change the outcome — without requiring them to query a dashboard

3. Capability Depth

What this question diagnoses: Whether the organization has moved from experimental AI to production-grade AI — specifically, whether it can deploy, monitor, and iterate on AI systems without rebuilding from scratch each time. The "Proof of Concept Factory" trap is a Level 2–3 phenomenon

AI exploration is underway — vendor demos attended, internal conversations happening — but nothing has been built or deployed internallyWe have run proofs of concept or sandbox experiments, but none have reached production or been validated against live business dataAt least one AI system is live in production, but it was built as a bespoke project — reuse, versioning, and fallback logic are not standartizedWe operate reusable AI components with documented model versioning, drift monitoring, fallback protocols, and defined ownership for each production systemOur AI engineering infrastructure supports multi-format inputs, self-monitoring outputs, and ships new capabilities into production in weeks — not quarters

4. Organizational Rewiring

What this question diagnoses: The 70% that determines whether AI delivers value — not the algorithm or the data, but whether people, roles, incentives, and operating processes have been redesigned around AI-augmented decision-making. Firms at L1–2 here will fail to scale regardless of technical maturity

AI is treated as an IT initiative — business unit leaders are not involved in use case design, deployment, or outcome measurementA small number of champions are using AI tools, but the majority of roles, workflows, and performance incentives are unchangedWe have begun role redesign and AI fluency training; at least one business function has a defined human-in-the-loop protocol for AI-assisted decisionsCross-functional AI squads — combining data engineers, domain experts, and process owners — are embedded and operating across multiple business unitsAI-augmented work is the organizational default: job design, hiring criteria, performance frameworks, and budget allocation all reflect an AI-first operating model

5. Trust Architecture

What this question diagnoses: Whether responsible AI is operational or merely documented — i.e., whether governance has teeth: measurable harm prevention, mandatory human gates, bias audit trails, and regulatory defensibility. Responsible AI is a competitive differentiator, not a compliance checkbox

AI ethics and governance have not been formally addressed — no policy, no risk inventory, no accountable ownerWe have an AI policy document, but it has not been operationalized — there are no enforcement mechanisms, audit logs, or defined escalation pathsRisk profiles and bias testing protocols exist for key AI systems; a human-in-the-loop review gate is defined for high-stakes decision outputsGovernance is outcome-based and measured: we track harm prevented, bias incidents resolved, and model accuracy against a defined threshold — not just policies writtenFull governance stack is in place: continuous model monitoring, proactive regulatory readiness (EU AI Act, GDPR, HIPAA), and every AI output is traceable to its source data and the logic that produced it

6. Value & Scale

What this question diagnoses: Whether the organization has broken out of the "AI spend without velocity" pattern — AI investment is rising, but measured business outcomes are not compounding. The test is not whether AI is in use; it is whether the economics are improving as it scales

We are investing in AI but cannot yet attribute any measurable change in revenue, cost, risk, or speed to a specific AI initiativeOne or two initiatives show directional promise, but ROI is not formally tracked — outcomes are reported anecdotally rather than against a baselineWe can quantify business impact for at least two AI systems in production, with a documented baseline, a defined metric, and a measurement cadenceSuccessful pilots have been replicated across teams or geographies using a standartized scaling playbook — the capability, not just the outcome, has been transferredAI ROI is governed like any capital investment: cost-per-outcome is tracked, improving, and reported to the board — and the portfolio is actively rebalanced based on that data

Disclaimer: The AI Maturity Scan is a directional diagnostic, not a formal audit or certification. Results are based on the inputs provided and are intended to help identify maturity patterns, capability gaps, and potential next steps. A deeper enterprise review may be required to validate findings and define a detailed transformation roadmap.