Skip to main content

Enrich your metadata

Use these workflows to enrich assets with descriptions, owners, tags, and business context from your AI client. Assets are often crawled into the catalog without being enriched—tables exist without descriptions, owners are unassigned, and classification tags are missing. These workflows cover bulk asset annotation, linking BI dashboards to business terms, propagating metadata through lineage, and creating glossary terms from how assets are already used across your catalog.

info

Make sure Atlan MCP is configured before running these workflows. For setup, see Set up Atlan MCP.

Enrich assets with your context​

Catalog assets are often crawled but never enriched—tables exist without descriptions, no one is assigned as owner, and tags are missing. This workflow lets you find those gaps and fill them in bulk using natural language, without opening the Atlan UI.

How it works
1
Search for assets by schema or keyword
semantic_search
2
Retrieve current metadata to identify gaps
get_asset
3
Update descriptions, owners, tags, and custom fields
update_assets
Example prompt
Find all tables in the orders schema that have no description and add descriptions based on their column names. Also set the data steward as owner.

Give your dashboards context​

Dashboards without owners, descriptions, or glossary term links don't surface in search and are hard to govern. This workflow connects your BI assets to the business context that makes them discoverable and trustworthy.

How it works
1
Search for dashboards by name or business area
semantic_search
2
Review enrichment gaps—missing owners, terms, descriptions
get_asset
3
Update with business terms, owners, and tags
update_assets
Example prompt
Find the Revenue Overview and Customer Churn dashboards in Tableau. Add the Finance team as owner and link them to the Revenue and Churn Rate glossary terms.

Propagate context through lineage​

When a source table has good descriptions and tags, every downstream view, table, and dashboard ideally inherits that context—but this doesn't happen automatically. This workflow traverses lineage and propagates metadata from source to consumer in one pass.

How it works
1
Get the lineage graph for a source asset
traverse_lineage
2
Walk downstream nodes—views, tables, dashboards
get_asset
3
Apply descriptions and tags from the source asset
update_assets
Example prompt
Take the description and PII tags from the customers_raw table and propagate them to all downstream tables and views in its lineage.

Grow your business glossary​

Business terms like 'churn', 'LTV', or 'active user' are used across dozens of assets but rarely have a canonical definition in the catalog. This workflow finds every asset referencing a concept and uses that context to bootstrap a glossary term.

How it works
1
Search for assets referencing a business concept
semantic_search
2
Draft a definition from how it's used across assets
get_asset
3
Create the term and link it to relevant assets
create_glossaries
Example prompt
Find all assets that reference customer lifetime value or LTV. Create a glossary term called Customer Lifetime Value with a definition based on how these assets use it, then link them to the term.