Here is a five-step plan to move your Purview Data Governance work into OneLake + Atlan before support ends on November 12, 2027. You will finish with your glossaries, domains, data products, and custom attributes in Atlan, checked against a plan your governance lead approved.
If your team runs data governance on Microsoft Purview, Microsoft’s November 12, 2026 announcement put an end date on everything your team built there. The glossary your stewards argued over, the owners you tracked down one domain at a time, and the data products other teams learned to trust took years to build. Losing that work now would mean rebuilding the definitions your AI agents depend on, at the moment you need them most.
Move only what your team built by hand, and let Atlan’s connectors rebuild everything Purview scanned. That covers metadata, the technical facts about each table and column, and lineage, the map of how data flows from source to report. Crawling both fresh means they start current, with your curation on top.
What is changing in Purview Data Governance, and what are you migrating?
Microsoft is retiring Purview Data Governance, and support ends on November 12, 2027. Microsoft Fabric with the OneLake catalog is now Microsoft’s data governance solution. Microsoft has published transition guidance, and your account team can help.
| What | Status | What you do |
|---|---|---|
| Purview Data Governance | Retiring, end of support November 12, 2027 | Migrate the governance work stored there |
| Sensitivity labels and Data Loss Prevention (DLP) | Not affected, stays with Microsoft Purview | Nothing. Keep them running |
| Data governance in Microsoft | Moves to Microsoft Fabric, with the OneLake catalog | Plan how the OneLake catalog and your wider catalog divide the work |
| Glossaries, domains, data products, custom attributes, owners, and hand-drawn lineage | Your decision | Choose where they go next |
The retirement explainer covers the announcement in full, and what Microsoft Purview is covers the product split. Your migration scope is your team’s curation, and that is the part worth protecting.
What do you need before you start?
You need two administrators, a named governance lead, and your main sources connected.
People:
- A governance lead who approves where every glossary, domain, and data product lands. One person, with the authority to say yes.
- A Microsoft Entra administrator who registers a read-only app for Migration Studio. Microsoft Entra ID is Microsoft’s identity service.
- A Purview administrator who grants that app two reader roles. Nobody grants Atlan an administrator role.
Systems:
- Migration Studio turned on. It is in private preview, and your Atlan account team turns it on.
- An Atlan tenant with the sources your glossary describes already connected. Atlan writes your governance onto the assets its connectors crawl, so connect sources first. The Fabric connector covers lakehouses, warehouses, semantic models, reports, Dataflows, and pipelines.
- An owner list you trust. Owners carry across when they match Atlan users, so check who has left. The owners and stewards guide covers what each role needs.
Decisions to make early:
- Whether expired terms and draft domains come across. To bring a draft domain across, publish it first.
- Whether Purview governance domains become Atlan domains, glossaries, or both. The data domain guide helps you choose.
Settle these before Step 3.
Step 1: Assess your Purview Data Governance estate
The assessment shows what you have and what moving it takes.
How to do it:
- Start today on your own: export your classic glossary terms to CSV, and list the domains and data products your teams use most.
- Request a migration assessment from Atlan, or from a Microsoft partner you already work with.
- Have your Entra administrator register the read-only app, and your Purview administrator grant it the two reader roles.
- Let Atlan read your estate. It writes nothing back.
- Read the plain-English report: what carries across, where your team decides, and a dated plan that finishes before November 12, 2027.
A business glossary that stays up to date starts with knowing which terms people use.
Done when: your governance lead has the assessment report, a dated plan, and a list of the decisions your team owns.
Step 2: Analyze the extract and list your decisions
Analysis turns the raw extract into a summary your team can act on before anything moves.
How to do it:
- Run Migration Studio’s extraction. Pre-flight checks confirm it can read your Data Map and Unified Catalog before anything runs.
- Run the analysis. It writes a plain-English summary of every glossary, term, domain, data product, and custom attribute it found.
- Work through the summary and list every decision your team needs to make.
Nothing moves until your team has read the summary.
Done when: every decision from the analysis has a named person to make it.
Step 3: Review where every object lands in Atlan
Review is where your governance lead shapes the result. Nothing is written until you approve.
How to do it:
- Choose which glossaries, categories, domains, and data products are in scope.
- Map each custom attribute to Atlan custom metadata or to a native Atlan field such as owners or description.
- Choose which classifications come across as Atlan tags. The data classification tools guide covers what to look for in tagging.
- Map each Purview source connection to the matching Atlan connection, and compare the two side by side.
- Decide how conflicts resolve when a term, domain, or product with the same name already exists in Atlan: keep, merge, or override.
- Map collections to domains, and redesign access as Atlan personas.
Export the plan so each domain owner can sign off offline.
Done when: your governance lead has approved the plan for every object type in scope.
Step 4: Move your governance and validate it against the plan
The move writes your approved governance into Atlan in two passes, then proves it landed.
How to do it:
- Create: Migration Studio writes glossaries, terms, domains, data products, custom metadata, tags, and hand-drawn lineage into Atlan.
- Validate: the reconciliation report reads Atlan back and puts what you planned next to what landed, object type by object type, with counts and coverage for each glossary.
- Enrich: Migration Studio writes descriptions, owners, tags, and custom metadata onto your crawled tables, columns, and Power BI reports.
- Fix links: references that still point at Purview Data Governance are rewritten to point at the Atlan assets.
- Resolve users: owners not yet in Atlan wait on a list until they join.
Every migrated object records where it came from: its Purview ID, URI, and collection.
Done when: the reconciliation report matches your approved plan, and your stewards are working in Atlan day to day.
Step 5: Extend governance past Fabric and into AI
Extending is what makes the move pay off, for your people and the AI agents your company is about to build.
How to do it:
- Connect: Atlan connects to more than 100 sources, including Snowflake, Databricks, SAP, Oracle, Teradata, and Tableau, with column-level lineage from source systems through Fabric into Power BI.
- Read from OneLake: Atlan’s Context Lakehouse keeps its context in OneLake as open Iceberg tables you own, so Fabric can query it directly, with no copies.
- Enrich: Context Agents write descriptions, build out your glossary, and tag each asset to its business domain, which automates stewardship. Your experts certify what goes live.
- Serve agents: Atlan’s MCP Server delivers certified context to Copilot Studio agents and any agent that supports the Model Context Protocol (MCP), an open standard for connecting AI agents to tools and data.
Done when: your top sources outside Fabric are connected, and at least one agent or Fabric workload reads governed context from Atlan.
What carries across from Purview Data Governance, and what does Atlan rebuild?
Your team’s curation carries across, and what Purview scanned, Atlan crawls fresh from your sources.
The reason it maps well: Purview’s Data Map implements the Apache Atlas API and type system, and Atlan runs on Apache Atlas. Apache Atlas is the open-source metadata model both products build on.
| What you’ve built | What happens in Atlan |
|---|---|
| Glossaries, categories, and terms, with synonyms, acronyms, related terms, status, and linked assets | Carry across |
| Governance domains and sub-domains | Become Atlan domains and glossaries |
| Data products, with status, sensitivity, owners, assets, terms, and documentation links | Carry across |
| Custom attributes, definitions and values | Carry across, typed to match |
| Critical data elements | Carry across as custom metadata |
| Objectives and key results | Become glossary terms linked to the data products they cover |
| Classifications on your assets, your own and Microsoft’s built-in ones | Become Atlan tags |
| Descriptions, display names, and owners | Carry across. Owners are matched to Atlan users, with experts and stewards recorded alongside |
| Lineage your team drew by hand | Carries across wherever both ends are in Atlan |
| Assets, technical metadata, and crawled lineage | Rebuilt by Atlan’s connectors, down to columns where the source exposes them |
| Data quality rules | Set up again in Atlan with your team |
| Sensitivity labels and DLP | Stay with Microsoft Purview |
Check terms against the glossary required fields before they cross, and keep data products whole.
How do you export glossary terms, classifications, and lineage from Purview yourself?
You can export parts of it yourself, and you should know how, even if you use Migration Studio. Microsoft documents these routes:
- Classic glossary terms: export to CSV from the Purview portal, per Microsoft’s import and export guide.
- Type definitions, including custom types and classifications: read them through the Data Map type REST API.
- Lineage: read entity lineage through the Data Map lineage REST API.
An export gives you files, and files are only half a migration. Someone still maps every field, matches owners, relinks terms to assets, and checks the result. A script and a spreadsheet can handle a small glossary. For domains, data products, and years of custom attributes, use a tool that keeps every decision reviewable. The migration best practices guide covers how to plan it.
What are the most common mistakes in a Purview Data Governance migration?
Each comes down to people or order.
- Moving the glossary before connecting sources. Terms need crawled assets to land on, so connect sources first.
- Treating owners as data. Owners carry across only when they match an Atlan user, so clean the list first.
- Copying scanned metadata. Assets and lineage that Purview scanned are a snapshot of the past. Let Atlan’s connectors crawl them fresh.
- Switching off Purview Data Governance too early. Run both systems until the reconciliation report matches your plan. A readiness checklist helps regulated teams set the bar.
- Moving sensitivity labels and DLP. They stay with Microsoft Purview, so leave them alone.
- Stopping at Fabric. Plan for Snowflake, Databricks, and SAP from the start, and check the Synapse to Fabric gaps if you are moving workloads too.
Get the order right: people and decisions first, sources second, governance third.
Is the OneLake catalog enough after Purview Data Governance?
The OneLake catalog comes with Fabric and governs data in Fabric, plus Azure and third-party sources through shortcuts, mirroring, and supported connectors. If your data lives almost entirely in Fabric, it may be all you need.
Atlan is for teams whose data goes further: sources outside Fabric governed where they run, column-level lineage across all of them, and governed context for Copilot Studio and any MCP agent. Compare the options Microsoft names in Purview Data Governance alternatives, and see what the platform includes in Microsoft Fabric governance.
If the data your glossary describes sits mostly outside Fabric, plan for both from day one.
Real stories from real customers: Governance at scale
"AI initiatives require more context than ever. Atlan's metadata lakehouse is configurable, intuitive, and able to scale to hundreds of millions of assets. As we're doing this, we're making life easier for data scientists and speeding up innovation."
— Andrew Reiskind, Chief Data Officer, Mastercard
"Context is the differentiator. Atlan gave our teams the shared vocabulary and lineage to move from reactive data management to proactive AI enablement."
— Kiran Panja, Managing Director, Cloud and Data Engineering, CME Group
Migrated governance becomes AI context once Context Agents keep building on it: as of October 2026, Context Agents have written more than 2 million descriptions across more than 200 organizations and saved more than 210,000 hours.
FAQs about migrating from Purview Data Governance
1. When does support for Purview Data Governance end?
Support for Purview Data Governance ends on November 12, 2027. Microsoft announced the retirement on November 12, 2026, and named Microsoft Fabric with the OneLake catalog as its data governance solution.
2. Is all of Microsoft Purview being retired?
No. Only Purview Data Governance is retiring. Purview Data Security and Compliance, including sensitivity labels and Data Loss Prevention, stays with Microsoft.
3. Can you export a Purview glossary to another catalog?
Yes. Classic glossary terms export to CSV from the Purview portal, and the Data Map REST API returns type definitions, entities, and lineage. Exported files still need mapping to the target catalog’s model, which is the work Migration Studio does for you in Atlan.
4. Can you run Purview Data Governance and Atlan side by side?
Yes. Both systems run side by side while you migrate. Migration Studio can extract a fresh snapshot whenever you need one, adding what is new without duplicating what already landed.
5. Do you need to move classic glossary terms into Unified Catalog first?
No. Unified Catalog is part of Purview Data Governance, which is retiring. Migration Studio reads both classic Data Map glossaries and Unified Catalog, and classic glossaries arrive in Atlan as their own glossaries.
6. Can you connect to Atlan through Microsoft Marketplace?
Yes. Connect to Atlan through Microsoft Marketplace. Atlan supports Microsoft Entra ID single sign-on.
7. What should regulated teams do first?
Talk to your Microsoft account team early. Microsoft asks customers with contractual obligations and in highly regulated industries to engage early, because those scenarios are assessed individually.
Sources
- Type definitions and how to create custom types, Microsoft Purview, Microsoft Learn
- Import and export glossary terms (classic), Microsoft Purview, Microsoft Learn
- Data Map type REST API, Microsoft Purview, Microsoft Learn
- Data Map lineage REST API, Microsoft Purview, Microsoft Learn
- Migrate classic glossary terms to Unified Catalog, Microsoft Purview, Microsoft Learn
- OneLake catalog overview, Microsoft Fabric, Microsoft Learn
- What does Atlan crawl from Microsoft Fabric, docs.atlan.com
- Migrate into Atlan: best practices, docs.atlan.com
- Remote MCP overview, docs.atlan.com
- Atlan on Microsoft Marketplace
- A next step in the evolution of the Microsoft Data Governance strategy, Microsoft Security Community Blog, November 12, 2026
