Understand what the Integrations Inventory governs

The Integrations Inventory governs every connection through which data or information moves between two entities — whether those entities are Applications, Databases, File Systems, Queues, APIs, Human actors, External Partners, or any other recognized Noun Type. A governed integration qualifies for a record when it represents a distinct, identifiable data or information flow with a named source, a named target, a describable payload, a specific technology, and a defined execution context (the environments involved). Every entry is a Noun Instance of the Integration Noun Type — a unique integration with its own Semantic ID, its own governance attributes, its own owner accountability, and its own risk profile.

An Integration is explicitly not an Application, a Technology, a Capability, or a Data Type — it is the governed connection between them. It is also known as a Data and Information Integration, reflecting that the Integrations Inventory supports both highly structured data payloads (relational records, JSON documents, CSV files, EDI transactions) and unstructured information payloads (PDF documents, images, audio recordings, binary files). The type of payload does not change whether something qualifies as a governed integration — only the presence of a distinct, identifiable data or information flow between two entities does. An informal manual file copy that moves sensitive data between systems is as much a governed integration as a fully automated API call.

The Integrations Inventory does not record every possible permutation of two entities communicating — it records every distinct, identifiable, governed data or information flow. Multiple integrations may exist between the same source and target (distinguished by their Integration Payload and Integration Type). A single bidirectional integration may be recorded as one record or split into two records when the two directions have meaningfully different governance characteristics. The governing principle is: one record per distinct governed flow, where distinct means it can be uniquely identified, has its own ownership, and can be independently monitored, governed, and retired.