Microsoft Purview Auto-Labeling for Legacy Office Files: Government Cloud Rollout Targeted for December 2026

The capability itself was announced earlier. What entry 578872 adds is a separate, later target for public sector tenants. That gap matters to agencies and contractors that often hold the oldest file shares in the building.

What's actually changing​

The roadmap description is short. Purview server-side auto-labeling will support legacy Office formats, and organizations can automatically evaluate and apply sensitivity labels to these files in SharePoint and OneDrive.

This is a change to automatic labeling, not manual labeling. Microsoft Learn treats two SharePoint and OneDrive features separately:

• Built-in labeling: users apply labels themselves in Office for the web or from a file's details pane.
• Auto-labeling policies: Purview inspects stored files and applies labels with no user involved.

The new support belongs to the second group. Microsoft Learn already says SharePoint and OneDrive can read and show a label that is already embedded in an uploaded .doc, .xls or .ppt file. Applying a label in Office for the web or SharePoint, however, is documented only for newer formats such as .docx, .xlsx and .pptx. So the change is not about SharePoint seeing old files for the first time. It lets Purview's policy engine open those files, check their contents against your conditions and apply a label.

Section summary: The change covers three legacy formats in two services, applied by policy on the server side. It is not described as a change to manual labeling.

Why this has been a blind spot​

Until now, auto-labeling simply skipped these formats. Office 365 for IT Pros said so plainly in 2025 when it discussed auto-label policies: older Office documents (.DOC, .XLS, and .PPT files) are unsupported, so some testing is needed to establish exactly what in the target content can be processed.

That gap hits hard in long-lived document libraries. Think of a procurement archive moved from an old file server, a spreadsheet from 2006 that still feeds a budget process, or a briefing deck nobody wants to open. These files can be sensitive while staying invisible to automated classification. Until now, the main fix was to convert them to modern formats first, which takes time and risks breaking old workflows.

Once this support arrives, conversion is no longer a requirement just to get these files labeled. Conversion can still make sense for other reasons. Microsoft has not said that migrating away from legacy formats is unnecessary.

Two roadmap IDs, two timelines​

Be precise about which entry you're reading:

The commercial entry, 574363, lists GA date: November CY2026 and the same month for preview. WindowsForum covered that entry earlier. Microsoft 365 Roadmap entry 574363, published on October 13, 2026, described the same .doc, .xls and .ppt support.

Microsoft later posted Message Center notice MC1530948 for the worldwide rollout. It gives a Public Preview: Beginning in early November 2026 and expected to complete by mid-November 2026 General Availability: Beginning in late November 2026 and expected to complete by late November 2026.

The descriptions match word for word, so 578872 looks like the government cloud version of the same feature. Microsoft has not said so directly, though. Government tenants should not assume the commercial schedule or rollout steps carry over exactly. Microsoft's roadmap also says its dates are estimates, that everything is subject to change, and that entries disappear once a feature reaches general availability, is cancelled or is postponed.

Section summary: Commercial tenants are looking at November 2026. GCC, GCC High and DoD tenants are looking at December 2026. Neither date is guaranteed.

The step that's easy to miss: existing policies won't pick this up on their own​

The most useful operational detail is in MC1530948, the commercial notice. It says existing policies need a minor edit and rerun to include these formats. New policies support them automatically. In other words, existing auto-labeling policies require a small policy edit and another policy run to restart the crawl and evaluate legacy Office files.

The notice gives an example of a small edit: add or remove a comment in the policy, save it and run the policy again. It also says the new formats work in both simulation mode and enforcement mode.

The government cloud entry, 578872, doesn't repeat any of this guidance. Treat it as the expected behavior based on the commercial notice, and check your own tenant's Message Center before relying on it. If the same rule applies to government clouds, a tenant that does nothing could have policies that keep ignoring legacy files even after general availability.

A practical checklist for when the feature reaches your tenant​

These steps follow the recommendations in MC1530948:

1. Confirm the prerequisite. Microsoft Learn's auto-labeling documentation says that to auto-label SharePoint and OneDrive files, you have enabled sensitivity labels for Office files in SharePoint and OneDrive. The documentation also notes that there's no check that you've enabled sensitivity labels for Office files in SharePoint and OneDrive when you configure auto-labeling, so confirm it yourself.
2. Review your policy inventory. Review your sensitivity labels and auto-labeling policies, including the SharePoint and OneDrive locations in scope.
3. Decide which policies should cover legacy files. Some may not need to. A policy aimed at an active project site may have no reason to scan a 2003-era archive.
4. Make a small edit and save it. Admins manage these policies in the Purview portal under Information Protection > Policies > Auto-labeling policies.
5. Rerun the policy so it starts a new crawl.
6. Run it in simulation first and review what it would label before you turn on enforcement.

Section summary: New policies should pick up the legacy formats automatically. Existing policies will probably need an edit and a rerun, and you should simulate before enforcing.

Things to plan for before you flip the switch​

Throughput limits. Microsoft Learn says a maximum of 500,000 SharePoint and OneDrive files can be automatically labeled per tenant per day, across all auto-labeling policies. The documentation also calls this a ceiling, not an expected rate. If your tenant holds millions of legacy files, a rerun that suddenly makes them eligible could take days, and it competes with your other policies for that daily limit. Bringing policies online in stages is the safer plan.

Licensing. Office 365 for IT Pros notes that anyone with access to a SharePoint Online site that's processed by an auto-label policy requires a license. Expanding which files a policy evaluates doesn't change that rule. If this feature tempts you to point policies at archive sites you used to ignore, check who has access to those sites first.

Existing limitations still apply. Microsoft Learn lists limits on SharePoint and OneDrive labeling that the roadmap entry doesn't mention. For example, the services can't read or apply labels on password-protected documents, and certain encryption setups block content processing. Nothing in 578872 says the new formats are exempt. Old password-protected .xls files are common in finance shares, so expect some files to fail.

Scope is narrow. The entry names three extensions and two storage services. It says nothing about other legacy formats such as templates, nothing about on-premises file shares, and nothing about Exchange. Don't assume more coverage than Microsoft has stated.

The bigger picture​

This is a small roadmap item that fixes a long-standing problem. Auto-labeling has always been only as good as its coverage, and the oldest files in many organizations were exactly the ones it skipped. Labels matter more as tools like data loss prevention and Copilot rely on them to decide what content can be shared or surfaced.

There is a fair complaint about the timing. Government tenants are probably the ones with the most decades-old documents, yet they again appear to be a month behind the commercial rollout. The good news is that the December estimate gives GCC, GCC High and DoD administrators time to prepare. They can use it to inventory policies, check the throughput budget and plan simulation runs, instead of finding out in January that their policies never rescanned the archive.

Summary​

• What: Purview server-side auto-labeling is adding support for .doc, .xls and .ppt files in SharePoint and OneDrive.
• Who: Roadmap 578872 covers GCC, GCC High and DoD tenants. The worldwide version is tracked as 574363 and MC1530948.
• When: General availability is estimated for December 2026 for government clouds and November 2026 for commercial tenants. Both dates can change.
• Action: Existing policies will probably need a small edit and a rerun to pick up the new formats. Simulate before you enforce, and check your own tenant's Message Center for the final details.