Skip to main content

AI Product Type (APT)

AI Product Type (APT) is the family-specific intrinsic classification of a qualified AI Product based on the provenance and construction structure of its productized intelligence.

APT answers one question:

Where does the Product's intelligence originate?

APT does not describe what the intelligence does, how autonomous it is, how it is executed, how risky it is, or whether the Product is physically embodied. Those are separate ontology dimensions.


APT01 — Primary AI Product​

A Primary AI Product is an AI Product whose productized intelligence is constituted without a material upstream AI Product dependency, directly from one or more AI Primitives together with governed data, knowledge, rules, tools, and controls.

Examples include:

  • a fraud-detection AI Product built from a registered classifier plus governed Data Products;
  • a symbolic reasoning AI Product built from a knowledge base, rules, and an inference engine;
  • an AI assistant built from a procured LLM Model Product, governed enterprise knowledge, policies, and guardrails.

A Primary AI Product may depend on many models, Data Products, knowledge sources, tools, or services. Multiple dependencies do not make it Derived unless the Product materially depends on an upstream AI Product intelligence proposition.


APT02 — Derived AI Product​

A Derived AI Product is an AI Product whose distinct intelligence proposition materially depends upon the governed intelligence capability or output of one or more upstream AI Products.

Example:

If Transaction Risk AI Product establishes a new governed intelligence proposition from those upstream intelligence capabilities, it is Derived.


Material Dependency​

A mere technical call to another AI Product is not enough to make a Product Derived.

The upstream AI Product dependency must be material to the resulting intelligence proposition. If the dependency is removed or replaced with a non-equivalent capability, the Product's essential intelligence proposition would materially change or cease to hold.

Examples that do not automatically establish Derived APT include:

  • logging to an AI-powered observability service;
  • using an AI Product only for optional text formatting;
  • routing requests to independent AI Products without creating a new intelligence proposition;
  • consuming a Model Product or generic inference endpoint as an AI Primitive.

Composite Derivation​

Composite is a derivation topology, not a peer AI Product Type.

A Derived AI Product may be:

  • single-upstream derived — materially derives its intelligence from one upstream AI Product; or
  • composite derived — materially composes multiple upstream AI Products into a new intelligence proposition.

This mirrors the distinction between provenance type and composition topology: composition describes how dependencies are combined; APT describes the origin of productized intelligence.


Model Product Dependency Does Not Imply Derived APT​

A downloadable or hosted learned model may be a governed Model Product and an AI Primitive without itself being an upstream AI Product.

Therefore:

Qwen Model Product
+
Enterprise Data / Knowledge / Policies
↓
Corporate Assistant AI Product

can still be classified:

APT = Primary

The distinction is essential for model vendors and organizations that build governed intelligence Products using procured models.


APT and AI Qualification​

APT is only meaningful after the Product qualifies for the AI Product Family.

The sequence is:

See AI Product Qualification & Evidence.


APT Is Version-specific​

APT may change across Product versions when the intelligence constitution changes.

For example:

A v1 → direct model dependency              → Primary
A v2 → upstream governed AI Product → Derived

or a Product may cease to qualify for the AI Product Family entirely if its final qualifying AI Primitive path disappears.

Historical versions retain their historical APT and qualification evidence.


What APT Is Not​

The following are not APT values:

TermCorrect ontology dimension
Predictive / GenerativeIntelligence Capability (IC)
AgenticAgency / Autonomy (AR) and Interaction / Execution Mode (MO)
SymbolicIntelligence mechanism / AI Primitive paradigm
HybridConstitution / mechanism composition characteristic
Physical / EmbodiedEmbodiment (EM)
R0–R4Risk Tier (RT)
Internal / ExternalConsumption Scope (CS)

Summary​

  • Primary means the intelligence is constituted without a material upstream AI Product dependency.
  • Derived means the new intelligence proposition materially depends on one or more upstream AI Products.
  • Composite describes Derived topology, not a separate peer APT.
  • Model Products, Data Products, rules, knowledge, and tools can be dependencies of a Primary AI Product.
  • APT is version-specific and supported by lineage and AI Qualification Evidence.

Principle: APT describes the provenance of productized intelligence, not its capability, technology, mode, agency, embodiment, or risk.