Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Note
This article applies to the preview version of AI capabilities in SharePoint (previously referred to as Knowledge Agent).
Important
AI in SharePoint capabilities require sites to be explicitly opted in to the AI in SharePoint public preview, or for the tenant to be opted in at the tenant level. Sites that are not opted in do not receive preview benefits.
Creating a document library with AI in SharePoint lets you go from a simple description to a well configured, metadata-rich document library. Instead of manually creating a library, defining columns, and setting up tagging rules, you describe what the library is for and optionally provide a sample file. AI in SharePoint then generates the library name, description, and a complete metadata schema—with columns that automatically tag new documents as they're added.
This feature is useful for organizations that need to quickly create well-organized document libraries, such as:
Finance and accounting: Invoice repositories, expense report libraries, financial statement archives.
Legal and compliance: Contract libraries, regulatory filing repositories, policy document collections.
Project management: Project deliverable libraries, product specs, milestone documentation.
Human resources: Employee records libraries, benefits documentation, training material collections.
How AI-powered library creation works
The library creation process uses AI to handle the setup that would otherwise require manual configuration by a site owner or content manager.
You describe the purpose: Provide a natural language description of what the library is for, for example, "I need a library for vendor invoices."
You provide a sample file (recommended): Upload an example or seed file that represents the type of content to store. AI uses this file to better understand the document structure and suggest relevant metadata fields. If you do not provide a sample file, AI generates the library configuration from your description alone.
AI generates the configuration: Based on your description and sample file, AI proposes a library name, description, and a metadata schema with columns tailored to your type of document.
AI creates and configures the library: The library is created, your sample file is copied to the library if one was provided, and AI-powered columns are configured. These columns use prompts to extract metadata from supported files added to the library.
Prerequisites
Users must have a Microsoft 365 Copilot license assigned.
The tenant and site must be opted in to AI in SharePoint preview.
Users must have permissions to create document libraries on the SharePoint site.
Steps for creating a library with AI
From the Site home page or Site contents page, select + New > Document library.
Select the AI-assisted creation process.
Describe the purpose of the library in the text field. For example:
"I need a library to store and organize vendor invoices."
"A library for marketing campaign assets and deliverables."
"Repository for employee onboarding documents."
Optionally, upload a sample file to help AI understand the structure of your documents. You can select a file from:
The current SharePoint site
A OneDrive location
Your local device
Review the AI-generated library configuration:
Library name: A descriptive name based on your input.
Library description: A summary of the library's purpose.
Select Create to confirm. AI creates the library, copies over your sample file, and configures the metadata columns.
AI-powered columns
Once the library is created, it includes AI-powered columns based on the metadata schema generated from your description and sample file. These columns use prompts to extract and populate metadata from supported files.
Automatic tagging: When documents are uploaded or dropped into the library, AI reads the content and automatically populates the column values. Users can review and update values as needed.
User control: The user can review the schema and the underlying prompts during creation and anytime after that.
Supported file types
AI-powered library creation uses the same underlying Autofill feature that reads document content and populates metadata values. This means the same file type compatibility applies to both the initial sample file you provide during library creation and to any documents added to the library afterward.