What does 'domain-oriented' mean in practice?

Instead of one flat, central glossary maintained by a small governance team, terms are grouped by business domain — Finance, Sales, Operations, Supply Chain, HR — and each domain has its own stewards who own the definitions. The same word can carry a domain-specific definition where that's the reality (e.g. 'Customer' in Sales vs. Finance), with the relationship made explicit instead of arbitrated away.

Where do the definitions come from?

From your own sources. Metagem reads SQL views and stored procedures, BI semantic layers (Power BI, Tableau, Looker), Confluence and SharePoint pages, data dictionaries, and existing catalog entries (Collibra, Purview, OpenMetadata). It extracts the definition the data actually implies, attaches the source it came from, and proposes a canonical version per domain.

How do you handle conflicting definitions across domains?

When the same term is used differently in two domains (Sales vs. Finance 'Customer', Operations vs. Commercial 'Order'), Metagem surfaces the conflict rather than hiding it. Each domain can keep its own definition, with the relationship and the context for each documented — so business users and AI agents know which one applies in which conversation.

How does this stay current once it's published?

Metagem continuously monitors your data landscape for schema changes, new terms, and definition drift. When something changes, the right domain steward gets a targeted nudge — not a generic email blast — with the proposed update and its source. This keeps the glossary accurate without a perpetual cleanup project.

Definitions everyone uses the same way

Everyone has a definition. Nobody has the same one.

Definitions sit in documents, SharePoint sites and people's heads. Every team keeps its own version, and the day you publish a glossary, it starts going out of date.

“Different terms mean the same thing, and the same term means different things.”
Engineering lead, European online retailer

search: active customer4 versions

.pdfSales handbook
“Ordered in the last 12 months.”
.docxCredit policy v3
“Has an open account and no credit block.”
.xlsxKPI definitions
“Ordered in the last 24 months.”
wikiData team wiki
“Any customer with status A.”

From four versions to one agreed definition.

One term, from scattered versions to an entry its owner has signed off.

1Every version in one place

Shadow definitions and synonyms, found.

Ordered in the last 12 monthsSales

Ordered in the last 24 monthsBI · conflicts

What you see

Where each version lives, who uses it, and where they contradict each other.

2Ready to approve

One proposed definition, with its sources.

Proposed glossary entry · Owner: Finance

Active customer

Invoiced at least once in the last 12 months, and not on credit hold.

Proposed fix

Keep one definition, map the synonyms to it, and retire the outdated versions.

3Signed off without a meeting

Agreed in Teams or by email.

New definition to approve: Active customer

✓ Approved in Teams by the finance controller

✓ Confirmed by email by the sales lead

The result

One agreed definition, linked to its owner. Everyone gets the same answer.

Definitions, with and without Metagem.

Today

Definitions spread over documents, SharePoint and people's heads.
Every team keeps its own version.
Agreeing one term takes long meetings and chasing people.
The glossary is out of date the day it's published.

With Metagem

Every shadow definition foundSynonyms included.
A proposed fix for every conflictOne definition with its sources, ready to approve.
Agreed where people already workIn Teams or by email. No workshop needed.
Kept up to dateWhen a source changes, the owner is asked to confirm.

Pick one domain. See how many versions of each term you have.

Let's talk