Skip to content

Business glossary

The Glossary is where your business vocabulary lives: one surface for browsing what your organization has modeled and for editing what those words mean. A term is a business entity (Customer, Order) or one of its attributes (Customer ID, order total), so the concepts and the facts about them sit together instead of on two separate pages.

Open it from Business > Glossary in the left navigation. The page is split in two: a tree of your terms on the left, and on the right a grid of the terms in whatever you selected. A term you open gets a panel beside the grid, and the grid stays where it is.

What counts as a term

Section titled “What counts as a term”

Every row in the grid carries a Type badge that tells you which kind of term you are looking at:

  • Entity: a business concept with no subtypes beneath it.
  • Parent: a business entity that has other entities nested under it.
  • Attribute: a fact about an entity. Attribute rows are indented under the entity they belong to and sorted by their position on that entity.

This page replaced the separate Business entities page, which is why entities and attributes now appear in one list. An existing bookmark still works: the old entities address redirects here and carries the whole query string with it, so a saved link to a specific entity opens that entity on this page.

Find a term

Section titled “Find a term”

Three controls narrow what you see, and they work together.

  • Group by Model or by Entity. Grouped by Model, the tree is a catalog: domains, then their models and submodels, then the entities that belong to each, with an Unassigned row at the end for entities that belong to no model yet. An entity that belongs to more than one model appears under each. Grouped by Entity, the group tiers disappear and you get entities alone, nested by their parent entity, each appearing exactly once.
  • The Attributes checkbox, beside that toggle, adds or removes the attribute leaves under each entity in the tree. It starts switched on, and your choice is remembered in your browser. It changes the tree only: attribute rows stay in the grid either way.
  • One search box, in the page header rather than over the tree, reading “Search terms, technical names and definitions”. It narrows the tree and the grid at the same time. In the tree, a branch survives if it matches or if something under it does, and matching branches expand so you can see what was found. In the grid, it matches a term’s name, its technical name, its parent term, and its definition text.

On the organization-wide page, a project scope select beside the search narrows everything to one project, or leaves it at All projects. A project or a workspace also has this same glossary on its own page, already pinned to what that project or workspace holds, so there is no scope select to set there.

Read the terms grid

Section titled “Read the terms grid”

Select anything in the tree and the right side lists its terms, with the selection’s name at the top and a count of how many entities and terms it holds. Select nothing and you get All terms.

These columns are shown by default:

  • Term: the entity’s or attribute’s name.
  • Type: Entity, Parent, or Attribute, as described above.
  • Parent term: for an attribute, the entity it belongs to.
  • Technical name: the name derived from your naming convention.
  • Definition: the authoritative prose for what the term means.
  • Status: Draft or Approved.
  • Sensitivity: the term’s value from the Sensitivity scheme.
  • Classifications: the remaining classification chips.
  • Mapped columns: how many source columns in the branch are matched to this term, so you can see which terms your data actually uses. The count covers the whole branch, so it can include columns outside the model you are looking at.

A small sparkle icon beside a value in the Term or Definition column is the AI badge: that value is still exactly what an accepted AI proposal wrote. Hover it to read “Accepted from an AI proposal”, then the confidence (for example “Confidence 85%”) and the AI’s reason when the proposal gave them, and finally “It stays marked until someone edits it.” The icon turns amber when the confidence is below 50%.

More columns are available from the column chooser, among them the data type and its length, precision, and scale, whether the attribute is an identifier or a business key, its position on the entity, its AI context and AI instructions, an entity’s usage and limitations, an attribute’s example values (read only in the grid, edited in the attribute’s panel), whether it was modeled by hand or derived, and how many models an entity belongs to. Attribute columns stay blank on entity rows rather than showing a placeholder, because an entity has no data type of its own.

Your column widths, filters, and sort order are remembered between visits.

Open one term

Section titled “Open one term”

Click a term’s name and it opens in a panel beside the grid. The panel carries the toolbar every detail panel shares, with Save and Discard changes for the edits it holds.

  • An entity opens the same editor you use from the canvas, with its Details, Attributes, and Relationships tabs. Edits on Details wait for Save; the Attributes grid saves each cell as you go.
  • An attribute opens its own panel, headed with the attribute name, an Attribute badge, and a subtitle carrying its technical name, data type, and whether it is an identifier or a business key. The panel edits Name, Technical name, the data type and its qualifiers, Ordinal, the Identifier and Business Key flags, Synonyms, Example values (up to ten, as chips), Transformation, Definition (the text the grid shows in its Definition column), AI context, and AI instructions, and holds your edits until you select Save in its toolbar; Discard changes puts the stored values back. The panel also states in words which entity and attribute this one references, or that it references none.

Press Escape, or select Close in the panel’s toolbar, to close the term; selecting something else in the tree closes it too. If the panel holds an unsaved edit, the Unsaved changes question comes first. With a relationship open from an entity, Escape goes back to the entity instead.

Edit a term in place

Section titled “Edit a term in place”

Two columns are editable directly in the grid, and the gesture is what tells them apart from opening a term:

  • A single click on a term’s name opens that term’s detail on the right.
  • A double click on the name starts renaming it where it sits.
  • A double click on the Definition cell opens a text box for the full definition.

Both save as you go. Editing needs the editor role, and like every other catalog change it happens on a feature branch: on main the grid is read only and tells you to switch to a branch. Viewers can read everything and change nothing.

Each row also carries a Draft action that asks AI to propose a definition for that term and sends the proposal to the Ask AI pane for you to review. A definition you apply from there carries the AI badge, with no confidence in its tooltip, because a drafted definition has no score.

Definition approval status

Section titled “Definition approval status”

A definition moves through two states:

  • Draft: written but not yet reviewed and approved.
  • Approved: an editor has approved it. Hover the Approved badge to see who approved it and on what date.

Editing the definition text of an approved term returns it to Draft and clears the approval stamp, whether you edit it here or on the term itself. That is deliberate: a row showing Approved is always one whose current wording was actually reviewed. The AI badge is a separate mark: it records who wrote the current text, not whether anyone approved it, so approving a definition leaves the badge in place, and editing the text clears both. Approving an entity’s definition happens on the entity, in the Governance section of its Details tab. See Data governance overview.

Approve definitions in bulk

Section titled “Approve definitions in bulk”

Tick rows in the grid and an action bar appears above it with Approve definition, Set business key, and Delete.

These actions apply to attributes only. If your selection includes entity rows, the bar acts on the attributes you ticked and says how many entities it skipped, so nothing happens silently. Select entity rows alone and the bar shows the count with no action to take: approve an entity’s definition from the entity itself.

Glossary versus governance status

Section titled “Glossary versus governance status”

The glossary and your governance status read from the same model but answer different questions:

  • The glossary is the vocabulary: where you look up what a term means, and where you write and approve that meaning.
  • Governance status is the completeness check: each entity and attribute shows what is still missing (no definition, an unapproved one, no steward, no mapped source) in its own panel.

Reach for the glossary when you want to understand or define a term; check governance status when you want to find and close gaps.

Reach the glossary from a model or a domain

Section titled “Reach the glossary from a model or a domain”

A model and a domain no longer carry a Glossary tab of their own. Instead, both detail pages offer an Open in Glossary action that brings you here with that model or domain already selected in the tree, so the grid lists exactly its terms and the tree is still there to step up or across from. Opened from inside a project or a workspace, the action keeps you in that scope rather than sending you out to the organization-wide page.

Section titled “Related guides”