Skip to main content

AI Product Maturity Checklist

Assess a qualified AI Product version against AIPCH01–AIPCH21. Your answers are saved locally (browser only).
PMDD measures maturity; it does not establish Productness, AI Product Family membership, APT, or risk tier.
Not Applicable items are excluded from the normalized score and require a rationale.

This AI Product Maturity Checklist may be used, shared, and adapted freely, including for commercial purposes, provided attribution is given to KaizenXOne (Founding Architect of AIPS).
“AI Product Maturity Checklist — KaizenXOne, AI Product Specification (AIPS), licensed under CC BY 4.0.”

🤖 AI Product Maturity Checklist

Not Assessable — confirm CURRENT AI Qualification Evidence first
Assess each applicable characteristic as Met, Partial, Not Met, or Not Applicable. N/A weight is removed from the denominator and requires an evidence-backed rationale. Mandatory governance obligations remain binding regardless of score.
#AIPCHCharacteristicChecklist ItemSatisfied Only When ...WeightStatusEvidence / N/A rationalePoints
1AIPCH01Domain-OwnedNamed AI Product Owner (AIPRO) Assigned... the AI Product is owned by a business domain that is accountable for the behavior and outcomes of the capability, with a named AIPRO responsible for its definition, governance, and lifecycle.8%
0
2AIPCH02DeployableIndependently Deployable AI Capability via Self-Service... the AI Product can be defined, composed, deployed, and managed end-to-end by the domain using self-service, with all technical complexity abstracted by the platform.6%
0
3AIPCH03Declarative SpecificationDefined via Structured Declarative AI Specification... the AI Product is fully represented through structured declarative artifacts (e.g., AIPDS), with no reliance on manual procedural definitions.5%
0
4AIPCH04Declarative Intelligence Definition & CompilationAI Product Is Defined and Compiled from Declarative Intent... the AI Product is defined through declarative intent, behavior, constraints, and dependencies, with implementation compiled by the platform; upstream Product composition is used only where the Product constitution requires it.5%
0
5AIPCH05DiscoverableRegistered in AI Marketplace or Catalog... the AI Product is published into a discovery surface where consumers can independently find, evaluate, and compare it based on declared capabilities and metadata.4%
0
6AIPCH06Self-DescribingRich Capability Metadata Embedded... the AI Product exposes sufficient metadata describing its behavior, inputs/outputs, constraints, dependencies, and usage context for independent understanding.4%
0
7AIPCH07Trustworthy (via Trust Signals)Emits Trust Signals... the AI Product exposes current, machine-consumable trust evidence covering qualification, dependencies, assurance, governance, reliability and applicable behavior-quality signals.5%
0
8AIPCH08ReusableReusable Across Use Cases or Domains... the AI Product is designed for reuse and is adopted across multiple contexts without requiring bespoke customization.4%
0
9AIPCH09SLA–SLO Backed & ObservableObservable with SLAs, SLOs, and Quality Guarantees... the AI Product is continuously observable with defined SLAs/SLOs and automated monitoring of performance and behavior.5%
0
10AIPCH10Compliant by DesignPolicy-as-Code and Governance Controls Enforced... governance policies (regulatory, ethical, risk) are bound at design time and enforced automatically at runtime via policy-as-code.5%
0
11AIPCH11AddressableProvides Well-Defined Output Ports... the AI Product exposes governed output ports (APIs, events, agent interfaces) independent of consumer experience layers.4%
0
12AIPCH12Semantically AlignedAligned to Enterprise and Domain Ontologies... the AI Product’s inputs, outputs, and behavior align with shared enterprise vocabularies and ontologies.4%
0
13AIPCH13Consumption-Driven IntentBuilt for Clear Business or Operational Purpose... the AI Product is explicitly defined against a business, operational, intelligence or action purpose traceable to consumer needs and intended use.4%
0
14AIPCH14Testable & VersionedVersion-Controlled and Behavior-Tested... the AI Product evolves through versioned releases with reproducible tests validating behavior and outcomes.5%
0
15AIPCH15Economically AccountableTracks Cost, Usage, and Value Metrics... cost, usage, contextual fitness, and realized value evidence are measurable across the lifecycle, with value claims carrying appropriate attribution confidence rather than equating adoption with value.4%
0
16AIPCH16Explainable & TransparentProvides Proportionate Explanation and Transparency... the AI Product provides explanation and transparency appropriate to its outputs, actions, consumers, risk and Governance Obligation Profile.5%
0
17AIPCH17Bias-Controlled & Fairness-MeasuredFairness Is Governed Where Applicable... where fairness obligations apply, relevant bias/fairness criteria are evidence-backed, evaluated at an appropriate cadence, and governed with mitigation or escalation; otherwise N/A requires rationale.5%
0
18AIPCH18Governed Evolution & AdaptationDeclares and Governs Its Evolution Posture... the AI Product has an explicit, governed evolution posture appropriate to its intelligence mechanism, including learningMode NONE where learning is not part of the Product constitution.4%
0
19AIPCH19Safe & Policy-Bound UsageUsage Boundaries and Safety Controls Enforced... the AI Product enforces defined usage boundaries, detects misuse, and applies safety controls automatically.5%
0
20AIPCH20Interoperable & ComposableSupports Composition with AI Products and Data Products... the AI Product interoperates through governed ports and can participate in declarative compositions without conflating ordinary orchestration with creation of a new Derived AI Product.4%
0
21AIPCH21Human-Centered OversightProvides Proportionate Human Oversight and Intervention... the AI Product implements the human oversight, intervention, suspension and accountability mechanisms required by its agency, embodiment, risk and Governance Obligation Profile; per-action human approval is not universally required.5%
0
APPLICABLE SCORE100/100 weight applicable0.0%

🎯 Prioritized Recommendations

High Priority — Not Met

  • AIPCH01 (8%): Assign an accountable AIPRO and register ownership, domain, and lifecycle metadata in the registry.
  • AIPCH02 (6%): Provide intent-driven self-service pipelines; ensure one-click publish, rollback, and environment parity.
  • AIPCH03 (5%): Adopt AIPDS; validate in CI; generate infrastructure, policies, and interfaces from spec.
  • AIPCH04 (5%): Enable intent-driven definition and compilation; eliminate unnecessary manual implementation translation; use declarative Product dependencies where applicable.
  • AIPCH07 (5%): Expose qualification/dependency/governance/assurance currentness, effective risk and mechanism-appropriate quality signals through AIPS-DPP and observability.
  • AIPCH09 (5%): Define SLIs/SLOs; implement dashboards, alerts, and runbooks; track behavioral metrics.
  • AIPCH10 (5%): Implement policy fabric (privacy, fairness, access); enforce via runtime controls and audit logs.
  • AIPCH14 (5%): Implement evaluation pipelines; version artifacts; enforce regression testing and changelogs.
  • AIPCH16 (5%): Expose explanation interfaces; provide reasoning traces or feature attribution where applicable.
  • AIPCH17 (5%): Determine fairness applicability from intended use and governance; where applicable define relevant metrics, evidence cadence, thresholds and mitigation.
  • AIPCH19 (5%): Define prohibited use; implement safety filters; log violations and overrides.
  • AIPCH21 (5%): Bind oversight to AR/EM/IE/risk obligations; provide required review, intervention, suspension, audit and escalation controls without imposing unnecessary HITL.
  • AIPCH05 (4%): Publish an eligible AI Product listing with current AQE/DPP and ontology metadata such as APT, IC, AR, EM, IE, risk and usage context.
  • AIPCH06 (4%): Publish capability descriptors; include IO schema, behavior summary, constraints, and dependency graph.
  • AIPCH08 (4%): Design stable interfaces; document reuse patterns; provide adapters and examples.
  • AIPCH11 (4%): Define standard ports; publish API contracts; separate product from UI/experience layers.
  • AIPCH12 (4%): Map schemas to ontology; use canonical terms and identifiers; link to semantic registry.
  • AIPCH13 (4%): Define use cases, KPIs, and acceptance criteria; document consumer scenarios.
  • AIPCH15 (4%): Combine FinOps with COVP usage, fitness and value-realization evidence; record attribution confidence and avoid treating adoption alone as value.
  • AIPCH18 (4%): Declare the evolution mechanism and controls: retraining, fine-tuning, prompt/rule/knowledge evolution, component replacement, dependency migration, or learningMode NONE as appropriate.
  • AIPCH20 (4%): Expose governed semantic ports and dependency contracts; distinguish federated/orchestrated use from Productized Composition that creates a new Derived AI Product.
ℹ️ Scoring & Qualification Guidelines
  • Met = full applicable weight.
  • Partial = 50% of applicable weight.
  • Not Met = 0 points while retaining the characteristic weight in the denominator.
  • Not Applicable = excluded from both points and denominator; rationale required.
  • PMDD precondition = CURRENT AQE for the assessed AI Product version.

≥ 80%: L2 High-Maturity AI Product · 50–79%: L1 Evolving AI Product · < 50%: L0 Emerging AI Product.

These are maturity bands only. Product qualification, AI Product Family membership, APT, R0–R4 risk, marketplace eligibility, and production authorization are governed separately.

🔐 Governance Tips
  • Reference AQE, AIPROD, AIPDS, AIPS-DPP, Governance Obligation Profile, Marketplace and COVP evidence where relevant.
  • Use N/A only when Product constitution or governance evidence supports non-applicability.
  • Do not use a high maturity score to override mandatory governance blockers or stale trust evidence.
  • Export JSON/CSV for auditable maturity history and evolution planning.