A data inventory is a list of the places where useful business records live. It describes each source, its contents, its owner and its limitations. It does not require moving everything into one system.
Include spreadsheets, accounting exports, order systems, shared folders and the knowledge kept in people's working notes. Record who understands each source, how often it changes and what it should be used for. An abandoned spreadsheet can look authoritative long after its owner has stopped maintaining it.
A small-business example
A small manufacturer wants to understand missed delivery dates. Orders live in the sales system, production dates are on a planning sheet, and delivery changes sit in a shared mailbox.
The inventory reveals that no single source records the whole journey. It also reveals that the planning sheet has only been maintained consistently since May. The first analysis can use that period openly instead of implying that a full year's evidence is available.
Try this
List five sources connected to one business question. For each, write: where it lives, what one record represents, who maintains it, when it was last checked and what it cannot tell you. Ask the people who use the records to correct your list.
Use Source of truth to decide which source governs each field, Freshness to describe timeliness, and Permission and purpose to decide what should enter an AI workflow.