Skip to main content

Context and Current Challenges

1. Background​

The concept of the product is pervasive in industrial, digital, and knowledge domains.
Yet, despite its universality, product representation remains domain-bound:

  • Data Products use declarative deployment (DPDS) and semantic descriptors (DPROD).
  • AI systems are partially covered by model interchange formats (ONNX, PMML), documentation frameworks (Model Cards, Datasheets for Datasets), and risk standards (NIST AI RMF, ISO/IEC 22989, EU AI Act).
  • Physical products rely on catalog ontologies such as GoodRelations, ETIM, or OntoSTEP.
  • Software and services use heterogeneous API catalogs, service descriptions, or proprietary registries.

The result is an ecosystem of siloed specifications, each addressing local needs but lacking a shared foundation.

2. Fragmentation​

Current product specifications are highly fragmented:

  • Each domain reinvents identity, ownership, interfaces, and lifecycle semantics.
  • Duplication leads to inconsistencies, especially when enterprises wish to integrate Data + AI + Services in a unified marketplace.
  • Interoperability across domains is costly, often requiring adapters or custom mappings.

3. Lack of Common Vocabulary​

There is no generalized ontology of “Product” that spans multiple domains.
Consequently:

  • A Data Product and an AI Product cannot be described using a single, shared schema.
  • Governance and compliance must be repeated for each domain-specific spec.
  • Cross-domain product discovery is inhibited by inconsistent vocabularies.

4. Governance and Trust Gaps​

Without a base product layer, enterprises face challenges in:

  • Expressing contracts and obligations (SLA, SLO, fairness, risk classification) consistently.
  • Attaching policies and metrics across diverse product categories.
  • Demonstrating lineage and provenance in mixed ecosystems (e.g., AI Products trained on Data Products).

This gap undermines trust, especially in regulated environments.

5. Marketplace Complexity​

As enterprises pursue unified marketplaces, fragmentation creates obstacles:

  • Multiple product catalogs must be maintained separately (data, AI, software).
  • Entitlement and access controls are duplicated across silos.
  • Users must learn inconsistent discovery models, reducing adoption and reuse.

A Base Product Specification (BPS) is needed to reduce this complexity.

6. Innovation Constraints​

Fragmentation also constrains innovation:

  • New product types (e.g., digital twins, simulation models, autonomous agents) must define product semantics from scratch.
  • Lack of a shared kernel slows experimentation and standardization.
  • Enterprises risk vendor lock-in where proprietary product catalogs dominate.

7. Strategic Challenge​

The absence of a unifying specification makes it difficult to achieve:

  • Coherent governance across Data, AI, and hybrid products.
  • Regulatory compliance when obligations cut across product boundaries.
  • Industry-wide interoperability, which requires shared semantics.

In summary:
The current landscape of product specifications is siloed, fragmented, and inconsistent.
This inhibits governance, trust, and marketplace integration. The Base Product Specification seeks to address these challenges by providing a neutral foundation from which domain-specific descriptors may extend.