Skip to main content

AI Product (AIPS) Model Reference

This reference documents the AI Product Specification (AIPS) core classes and properties in a structured format.
Each property is presented with a formal table (Identifier, Label, Domain, Range/Type) followed by plain-language explanation and examples.


AIProduct

Canonical class representing a true AI Product.
Characteristics: deployable, self-describable, governed, discoverable, consumption-driven.
If these characteristics are missing, the artifact may be only an AI Asset, not a full AI Product.


aips:label

PropertyValue
Identifieraips:label
LabelAI Product label
Domainaips:AIProduct
Range/Typexsd:string

Definition:
A concise, human-readable name of the AI Product.
Should be unique within a discovery/catalog system.

Example: "Multilingual Legal Summarizer"


aips:description

PropertyValue
Identifieraips:description
LabelAI Product description
Domainaips:AIProduct
Range/Typexsd:string

Definition:
A plain-language explanation of the AI Product, outlining its purpose, scope, and core functionality.
This is the first field most consumers will see in a catalog.

Example: "Summarizes legal contracts into plain-language abstracts for non-specialist users."


aips:aiProductOwner

PropertyValue
Identifieraips:aiProductOwner
LabelAI Product owner
Domainaips:AIProduct
Range/Typefoaf:Agent or xsd:anyURI

Definition:
The accountable individual, team, or organization responsible for the AI Product.
Critical for governance, compliance, and trust.

Example: "https://example.org/orgs/kivanura-foundation"


aips:capabilityType

PropertyValue
Identifieraips:capabilityType
LabelAI Product capability type
Domainaips:AIProduct
Range/Typeaips:CapabilityType

Definition:
Declares the primary AI capability of the product.
Values are drawn from the controlled vocabulary: Predictive, Generative, Agentic, Symbolic, Hybrid.

Example: aips:Generative


aips:inputPort

PropertyValue
Identifieraips:inputPort
LabelInput port
Domainaips:AIProduct
Range/Typebps:DataProduct or xsd:anyURI

Definition:
Specifies dependencies or inputs required by the AI Product.
Inputs may include data products, signals, APIs, or models.

Example: "urn:dataproduct:contracts-dataset:v2.0.0"


aips:outputPort

PropertyValue
Identifieraips:outputPort
LabelOutput port
Domainaips:AIProduct
Range/Typebps:DataProduct or xsd:anyURI

Definition:
Defines the outputs or capabilities the AI Product provides.
Outputs may include structured data, generated text, embeddings, APIs, or artifacts.

Example: "urn:aiprod:legal-summarizer-api:v1"


aips:purpose

PropertyValue
Identifieraips:purpose
LabelPurpose/intent
Domainaips:AIProduct
Range/Typexsd:string

Definition:
Captures the intended goal or business purpose of the AI Product.
Supports governance, ethical assessment, and alignment with organizational strategy.

Example: "Enable non-lawyers to understand key contractual obligations quickly."


aips:governancePolicy

PropertyValue
Identifieraips:governancePolicy
LabelGovernance policy reference
Domainaips:AIProduct
Range/Typexsd:anyURI

Definition:
Links to applicable governance artifacts such as model cards, risk assessments, audits, or compliance reports.

Example: "https://example.org/policies/aiprod-legal-summarizer-model-card"


aips:lifecycleStage

PropertyValue
Identifieraips:lifecycleStage
LabelLifecycle stage
Domainaips:AIProduct
Range/Typeaips:LifecycleStage

Definition:
Indicates the current stage of maturity of the AI Product.
Allowed values: Experimental, Beta, Production, Deprecated, Retired.

Example: aips:Production


aips:riskClassification

PropertyValue
Identifieraips:riskClassification
LabelRisk classification
Domainaips:AIProduct
Range/Typeaips:RiskClassification

Definition:
Categorizes the AI Product’s risk profile, aligned with frameworks like the EU AI Act.
Allowed values: MinimalRisk, LimitedRisk, HighRisk, UnacceptableRisk.

Example: aips:HighRisk


aips:monitoringMetadata

PropertyValue
Identifieraips:monitoringMetadata
LabelMonitoring metadata reference
Domainaips:AIProduct
Range/Typexsd:anyURI or xsd:string

Definition:
Points to observability, monitoring, or logging dashboards that track product health, fairness, drift, and reliability.

Example: "https://grafana.example.org/dashboards/aiprod-legal-summarizer"



AIService

Represents service-level access to AI Product functions (inference, retraining, orchestration).
An AIService makes an AI Product usable via protocols and endpoints.


aips:isAccessServiceOf

PropertyValue
Identifieraips:isAccessServiceOf
LabelAccess service binding
Domainaips:AIService
Range/Typeaips:AIProduct

Definition:
Binds the service instance to its associated AI Product.
Ensures discoverability and traceability.

Example: "urn:aiprod:legal-summarizer:v1"


aips:protocol

PropertyValue
Identifieraips:protocol
LabelAccess protocol
Domainaips:AIService
Range/Typexsd:string

Definition:
Specifies the communication protocol used for service access.
Common values: HTTP, gRPC, MQTT, custom.

Example: "HTTP"


aips:securitySchemaType

PropertyValue
Identifieraips:securitySchemaType
LabelSecurity schema type
Domainaips:AIService
Range/Typexsd:string

Definition:
Indicates the security model governing the service.
Values may include OAuth2, APIKey, mTLS, ZeroTrust.

Example: "OAuth2"


aips:endpointURL

PropertyValue
Identifieraips:endpointURL
LabelEndpoint URL
Domainaips:AIService
Range/Typexsd:anyURI

Definition:
Provides the URI where the AI Product service is accessible.

Example: "https://api.example.org/legal-summarizer/v1"


aips:endpointDescription

PropertyValue
Identifieraips:endpointDescription
LabelEndpoint description
Domainaips:AIService
Range/Typexsd:string or xsd:anyURI

Definition:
Human- or machine-readable description of the endpoint.
May point to an OpenAPI spec, GraphQL schema, or inline text.

Example: "https://api.example.org/legal-summarizer/openapi.yaml"


Distribution

Defines how an AI Product is packaged and distributed to consumers.


aips:accessService

PropertyValue
Identifieraips:accessService
LabelAccess service reference
Domainaips:Distribution
Range/Typeaips:AIService

Definition:
Identifies the service associated with a given distribution.

Example: "urn:aiservice:legal-summarizer-api"


aips:conformsTo

PropertyValue
Identifieraips:conformsTo
LabelConformance declaration
Domainaips:Distribution
Range/Typexsd:anyURI

Definition:
Indicates adherence to standards, regulations, or best practices.
Examples: ISO/IEC 42001, EU AI Act, internal audit schema.

Example: "https://example.org/standards/iso-42001"


aips:isDistributionOf

PropertyValue
Identifieraips:isDistributionOf
LabelDistribution source reference
Domainaips:Distribution
Range/Typeaips:AIProduct

Definition:
Links a distribution back to the AI Product it represents.

Example: "urn:aiprod:legal-summarizer:v1"


aips:format

PropertyValue
Identifieraips:format
LabelDistribution format
Domainaips:Distribution
Range/Typexsd:string

Definition:
Specifies the packaging format.
Values may include: ContainerImage, APIEndpoint, ModelWeights, AgentBundle.

Example: "ContainerImage"


KnowledgeSource

Represents datasets or knowledge sources used by the AI Product.
Crucial for provenance, reproducibility, and trust.


aips:sourceLabel

PropertyValue
Identifieraips:sourceLabel
LabelSource label
Domainaips:KnowledgeSource
Range/Typexsd:string

Definition:
Human-readable name of the dataset or knowledge source.

Example: "Contracts Corpus v2"


aips:sourceDescription

PropertyValue
Identifieraips:sourceDescription
LabelSource description
Domainaips:KnowledgeSource
Range/Typexsd:string

Definition:
Summary of the dataset’s purpose, scope, and content.

Example: "Curated corpus of 10,000 anonymized contracts from 2015–2020."


aips:sourceType

PropertyValue
Identifieraips:sourceType
LabelSource type
Domainaips:KnowledgeSource
Range/Typexsd:string (enumerated)

Definition:
Indicates how the dataset was obtained: Synthetic, Curated, Observational, UserProvided.

Example: "Curated"


aips:provenance

PropertyValue
Identifieraips:provenance
LabelProvenance
Domainaips:KnowledgeSource
Range/Typexsd:string or xsd:anyURI

Definition:
Captures the origin, lineage, and context of the dataset.
May reference external registries or data product catalogs.

Example: "sourced-from: https://datamarket.example.org/contracts"


aips:qualityMetrics

PropertyValue
Identifieraips:qualityMetrics
LabelQuality metrics
Domainaips:KnowledgeSource
Range/Typexsd:string or xsd:anyURI

Definition:
Documents quality, coverage, fairness, and drift metrics associated with the dataset.

Example: "Accuracy: 96%, Coverage: 90%, Bias Score: <0.05"


AgenticExtension (Optional)

Extends AIPS to model AI Agents, multi-agent systems, or emerging AGI/ASI artifacts.
This section is forward-looking and may evolve as the field matures.


aips:roleDefinition

PropertyValue
Identifieraips:roleDefinition
LabelAgent role definition
Domainaips:AgenticExtension
Range/Typexsd:string

Definition:
Defines the functional role of the agent in the system.

Example: "Contract negotiation bot"


aips:interactionProtocol

PropertyValue
Identifieraips:interactionProtocol
LabelInteraction protocol
Domainaips:AgenticExtension
Range/Typexsd:string

Definition:
Specifies the communication or messaging protocol used by the agent.
Examples: FIPA-ACL, LangGraph, Custom.

Example: "FIPA-ACL"


aips:coordinationMechanism

PropertyValue
Identifieraips:coordinationMechanism
LabelCoordination mechanism
Domainaips:AgenticExtension
Range/Typexsd:string

Definition:
Describes how the agent coordinates with peers or systems.
Modes include: Centralized, Swarm, Emergent.

Example: "Swarm"