Your supplier agreements are all in the right SharePoint library. But to find out when it renews, who signed it or whether notice is due, someone still has to open the file and read it. Multiply that by hundreds of documents and a well-organised library becomes a slow way to answer simple questions.
Autofill columns bring selected information, such as renewal dates, owners and approval statuses, into the library as searchable, filterable metadata. In this guide, you’ll learn how to create Autofill columns, test instructions and check the results before applying the approach more widely.
This is the first part of a three-part series. Part 2 will use the extracted information to create alerts with SharePoint rules, and Part 3 will go further if you need more specific actions and conditions to be met using Power Automate.
Image - The same contract library after Autofill has surfaced fields including parties, renewal dates, summaries and risk information.
Licensing
SharePoint permissions
Supported file types
Autofill columns use AI to extract information from documents and store it as metadata in a SharePoint column. The latest enhancement to SharePoint now enables a Gen AI prompt to be used in the column. Think of a prompt you might use in Copilot to extract certain information from a document. That prompt can be applied at scale across the document library when used in an Autofill column.
Each column combines:
For example, Autofill can identify document owners, extract key dates, classify statuses or generate summaries. The original files remain unchanged, but the information inside them becomes easier to search, filter and organise across the library.
Once configured, Autofill can process documents already held in the library as well as new files added later.
Before creating an Autofill column, identify a piece of information people regularly need to find inside documents. The best candidates are details that would make the library easier to search, filter or organise, such as an owner, date, category or status.
For example:
Start with a single column rather than trying to extract everything at once. This makes it easier to test the results, refine the Autofill instruction and confirm the information is genuinely useful before expanding further.
Start in the SharePoint document library where you want to surface information from documents.
Select Add column from the top of the library and choose the column type that best matches the information you want to extract. More about choosing the right column and instructions below.
For example:
Once you've created the column, turn on Autofill, open the settings and describe the information you want SharePoint to identify.
The clearer the instruction, the more useful the result is likely to be.
An example prompt:
Extract the contract renewal date from this document. If more than one date is present, identify the renewal date rather than the contract start date or signature date.
Before saving the column, use the built-in testing option against a representative document.
Don't stop at the first successful result. Test against several documents, particularly those with different formats, layouts or terminology.
A prompt that works perfectly for one document can produce inconsistent results across an entire library.
Once you're satisfied with the results, save the column and allow SharePoint to process the files.
The column type isn't just about how the information looks. It affects what you can do with the result later.
For example:
The question to ask is:
Once this information is visible, what do people need to do with it? If the answer is "nothing beyond reading it", a simple text field may be enough.
If the information will later drive notifications, approvals or automation, a more structured column type is usually the better choice.
Think of the instruction as a brief for a new colleague.
Explain:
For example:
Summarise this document in no more than 40 words. Use plain English and do not include a heading.
Or:
Identify the contract renewal date. Do not return the signature date or contract start date.
Use the testing option before saving the column and review the output carefully.
Microsoft recommends reviewing AI-generated results rather than assuming they are correct. The original document should remain the source of truth for important business decisions.
You can create each Autofill column individually, but Copilot can also help plan a broader set of columns for the library.
From inside the document library, Copilot can review the available documents and suggest:
This can be a helpful starting point if you know what you want the library to achieve but are unsure how the individual columns should be structured.
To use Copilot to create your Autofill columns, click on the Copilot icon in the SharePoint site and enter:
Review this library and suggest a set of suitable Autofill columns to use.
Copilot’s suggestions should still be treated as a draft. Before creating the columns, consider whether each one has a clear purpose, whether the proposed column type suits the intended use and whether the instruction reflects the documents held in the library.
You do not need to accept every suggestion. Start with the columns that will provide the clearest benefit, then expand the library once you know the results are useful and consistent.
After saving an Autofill column, allow SharePoint time to process the documents and then refresh the library. The results may not appear immediately.
Review the completed column across several files rather than relying on a single example. Check that the information is:
If a result is incomplete or inconsistent, return to the Autofill setup, adjust the instruction and test it again.
Autofill columns apply across the document library, so begin in a test library or with a controlled collection of files before introducing them more widely. This is particularly important if the library contains several unrelated types of documents.
AI-generated metadata should also remain open to review. Keep the original document accessible so people can check important information against its source, particularly where the data may influence approvals, deadlines or other business decisions.
Before extending the setup, select a small sample that includes normal documents and difficult cases. For each file, compare the generated value with its source. Record whether the value is correct, blank when it should be blank and formatted consistently.
Autofill sits inside Copilot in SharePoint, which is still in preview. For now, it only appears for people with a Microsoft 365 Copilot licence, and there is no pay-as-you-go setup to complete, so testing it costs you nothing extra (Set up and manage autofill columns).
That changes once it leaves preview. Autofill is then billed as a pay-as-you-go document processing service through a linked Azure subscription, currently $0.005 per transaction. A transaction is a page for Word, PDF or TIFF files, a sheet for Excel and a slide for PowerPoint (Pay-as-you-go pricing for document processing).
The cost to watch is reprocessing. Every time a document is updated and Autofill is run against it again, those pages are charged again. A 30-page contract reissued three times is 120 pages of processing, not 30. That makes Autofill a better fit for semi-static libraries such as contracts, policies or documentation than for busy working libraries where files change every day.
You have now surfaced key information that people can use without opening every document. More importantly, you have tested whether the instruction works across the library rather than trusting a single result.
People can use this structured information to search, filter and organise documents more effectively. It can also support clearer SharePoint views and provide the foundation for notifications and automated workflows.
In Part 2, we’ll use the column to create a SharePoint rule that alerts the right person when a document needs attention.
If you want to learn more about what SharePoint can do for you, read our page about SharePoint.
Autofill columns are a useful starting point, but the right setup depends on your documents, your teams and what needs to happen next. We can help you identify the information worth surfacing and turn it into a SharePoint experience that is easier to search, manage and automate.
Book a call to explore what this could look like for your organisation.