SharePoint Autofill Columns Just Got WAY Better (Now with Managed Metadata!)

This new SharePoint update is awesome. Autofill columns in Microsoft 365 now support Managed Metadata, making it easier than ever to automatically tag and classify your content using company terms.

If your information governance team blocked Autofill in the past, this changes everything. You can now use your term sets directly in Autofill columns for consistent, high-quality metadata across your libraries.

Watch the demo to see how it works and what limitations (like the 100-term cap) you should know about.

Video SummaryAutofill Columns Just Got Smarter: Microsoft 365’s autofill columns now support managed metadata — finally letting you connect AI tagging with your organization’s term sets.No More Manual Tagging Hassles: The feature recognizes preferred terms and synonyms automatically, so “Fabricam Inc.” and “Fabricam” are treated as the same clean value.New vs. Old: Earlier AI tagging relied on traditional machine learning. The new version taps into modern AI, giving you more accurate and flexible tagging.Hands-on Example: In the demo, autofill processed invoices, tagged existing clients correctly, and even added new terms like “Pluralsight Inc.” directly into the term store.Know the Limits: Currently, it supports up to 100 terms per set — ideal for departments or categories — and it’s available to try free in your tenant until the end of 2025.

For more information, read the transcript blog below, or watch the video above!

Transcript

If you want to use autofill columns to better classify your data in Microsoft 365, but your information governance architect wouldn’t allow it before, good news! Autofill columns now support managed metadata. In a recent update to the documentation, Microsoft announced that autofill columns now work with managed metadata using term sets containing up to 100 terms. When a preferred term or synonym matches, autofill will automatically respond with the preferred term.

Now, you might be wondering, “Wait, Vlad — isn’t that something we could already do?” That’s a great question, and it’s easy to get confused here. For the past two or three years, Microsoft has had a machine learning–based option that could automatically add managed metadata to SharePoint Online. If you had a pay-as-you-go set up, whenever you configured a managed metadata column, you could choose to automatically tag documents with terms. However, that earlier version had some limitations — for example, a maximum of five columns per library — and it wasn’t based on the latest AI technology. It relied on traditional machine learning, not large language models. That’s why, in Microsoft’s most recent presentations on document processing in Microsoft 365, this older approach was categorized under the “previous generation” of tools.

Many of my clients were hesitant to invest in that older solution, even though autofill columns worked well. The big issue was that we couldn’t ensure it used company-approved terms. That’s exactly what changes with this update — now, autofill columns can use managed metadata directly.

Let’s see it in action. In my demo, I’m in a SharePoint site where I’ll create a new document library called “Invoice Test.” Once it’s created, I’ll upload a sample document. In my SharePoint Online admin center, I already have a term set for clients in the term store. I want to make sure that when I generate invoice reports, all client names are consistent. So, I’ll add a managed metadata column called “Client.” When setting it up, you’ll notice two options: “Automatically tag documents with terms” and “Autofill.” The first option is the old taxonomy tagging method, and when you enable autofill, it automatically disables the old one — you can’t use both at once.

Next, I’ll select my “Clients” term set and set up a simple prompt: “Please extract the full name of the client from the invoice.” When I test it, it should identify “Fabricam,” which is one of my existing terms. What’s great is that managed metadata supports synonyms. For example, I can add “Fabricam Inc.” as a synonym for “Fabricam.” This means that whether a document says “Fabricam,” “Fabricam Inc.,” or a similar variation, autofill will always tag it correctly as “Fabricam,” keeping the metadata consistent across the library.

If your term set is open, autofill can also create new values automatically — you can decide whether to allow that. In my example, since the term set is open, it’s enabled by default. I can choose to disable it if I want more control. Once that’s set up, I’ll process the document through autofill and upload a few more test files.

For instance, I’ve uploaded a PDF labelled “Pluralsight” and another labelled “Pluralsight Inc.” — neither of which exists in my term store yet. After processing, autofill correctly tagged “Fabricam” and automatically added “Pluralsight Inc.” as a new term in the “Clients” term set. That’s impressive — it means autofill isn’t just applying metadata, it’s enriching your taxonomy dynamically.

You can verify the results by refreshing the term store. In my demo, I can see that “Pluralsight Inc.” and “Woodgrove Bank” were both added automatically. The library now shows all three clients tagged correctly, using managed metadata.

One thing to keep in mind is that the current limit for managed metadata in autofill columns is 100 terms. For many scenarios — such as departments or fiscal years — that’s more than enough, but larger organizations might find it restrictive for now. I’m confident Microsoft will continue improving and expanding this capability.

Finally, if you’d like to explore this yourself, you can try autofill columns for free in your Microsoft 365 tenant until the end of 2025. And if you want to stay up to date with the latest document processing and AI features in Microsoft 365, make sure to subscribe to Vlad’s Tech on YouTube — there’s always something new to learn.