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 Applicableitems 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| # | AIPCH | Characteristic | Checklist Item | Satisfied Only When ... | Weight | Status | Evidence / N/A rationale | Points |
|---|---|---|---|---|---|---|---|---|
| 1 | AIPCH01 | Domain-Owned | Named 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 | ||
| 2 | AIPCH02 | Deployable | Independently 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 | ||
| 3 | AIPCH03 | Declarative Specification | Defined 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 | ||
| 4 | AIPCH04 | Declarative Intelligence Definition & Compilation | AI 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 | ||
| 5 | AIPCH05 | Discoverable | Registered 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 | ||
| 6 | AIPCH06 | Self-Describing | Rich Capability Metadata Embedded | ... the AI Product exposes sufficient metadata describing its behavior, inputs/outputs, constraints, dependencies, and usage context for independent understanding. | 4% | 0 | ||
| 7 | AIPCH07 | Trustworthy (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 | ||
| 8 | AIPCH08 | Reusable | Reusable Across Use Cases or Domains | ... the AI Product is designed for reuse and is adopted across multiple contexts without requiring bespoke customization. | 4% | 0 | ||
| 9 | AIPCH09 | SLA–SLO Backed & Observable | Observable 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 | ||
| 10 | AIPCH10 | Compliant by Design | Policy-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 | ||
| 11 | AIPCH11 | Addressable | Provides Well-Defined Output Ports | ... the AI Product exposes governed output ports (APIs, events, agent interfaces) independent of consumer experience layers. | 4% | 0 | ||
| 12 | AIPCH12 | Semantically Aligned | Aligned to Enterprise and Domain Ontologies | ... the AI Product’s inputs, outputs, and behavior align with shared enterprise vocabularies and ontologies. | 4% | 0 | ||
| 13 | AIPCH13 | Consumption-Driven Intent | Built 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 | ||
| 14 | AIPCH14 | Testable & Versioned | Version-Controlled and Behavior-Tested | ... the AI Product evolves through versioned releases with reproducible tests validating behavior and outcomes. | 5% | 0 | ||
| 15 | AIPCH15 | Economically Accountable | Tracks 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 | ||
| 16 | AIPCH16 | Explainable & Transparent | Provides Proportionate Explanation and Transparency | ... the AI Product provides explanation and transparency appropriate to its outputs, actions, consumers, risk and Governance Obligation Profile. | 5% | 0 | ||
| 17 | AIPCH17 | Bias-Controlled & Fairness-Measured | Fairness 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 | ||
| 18 | AIPCH18 | Governed Evolution & Adaptation | Declares 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 | ||
| 19 | AIPCH19 | Safe & Policy-Bound Usage | Usage Boundaries and Safety Controls Enforced | ... the AI Product enforces defined usage boundaries, detects misuse, and applies safety controls automatically. | 5% | 0 | ||
| 20 | AIPCH20 | Interoperable & Composable | Supports 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 | ||
| 21 | AIPCH21 | Human-Centered Oversight | Provides 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 SCORE | 100/100 weight applicable | 0.0% | ||||||
🎯 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.