Collections overview in Knowledge Base

Collections give teams a simple way to organize knowledge, control access, and power AI-driven interactions from one place.

Collections are the foundation of how Knowledge Base organizes, manages, and activates your knowledge. A collection brings related documents, web sources, and shared context together so teams can work with information as a single, coherent asset instead of scattered files.

Whether you are preparing internal research, supporting customer-facing teams, or enabling AI-powered insights across the organization, collections make it easy to structure knowledge in a way that stays current, secure, and usable.

What collections enable

Collections are designed to reduce friction between information and insight. They let teams focus on outcomes instead of file management.

With collections, you can:

• Group related documents and sources into a single, queryable knowledge set.
• Keep content fresh through refresh and re-indexing controls.
• Share knowledge securely with individuals, teams, or the entire organization.
• Protect high-value collections from accidental deletion or change.
• See which documents are actually being used and referenced.

Collections act as the bridge between your content and the AI Assistant, allowing questions, search, and summaries to work against a curated and trusted knowledge base. 

Every response is grounded in your selected sources and includes citations back to the exact content it's based on, so you can verify answers before relying on them.

Supported content types

Collections support a broad range of commonly used document and web formats, allowing teams to work with existing content without conversion or restructuring.

Supported content includes:

• Adobe PDF documents.
• Microsoft Word, Excel, and PowerPoint files.
• Common text-based formats such as TXT and CSV.
• Email and web-based content, including public websites.
• Documents synced from supported storage integrations.
• Documents synced from supported storage integrations, such as SharePoint and Google Drive. 

PDF and Microsoft Word documents can also include images and figures, such as graphs and tables. Knowledge Base processes supported visual content along with the document text, allowing that information to contribute to search and AI-generated responses.

This flexibility allows collections to reflect how teams already store and work with information, while keeping content accessible for search, analysis, and AI-powered interactions.

Refer to the system requirements for a comprehensive list.

Built for easy management at scale

Collections are intentionally lightweight to create and simple to manage, even as they grow.

From a single interface, you can:

• Create and organize collections around projects, topics, or teams.
• Add documents from local files, connected storage, or public web sources.
• Update content as information changes, without rebuilding the collection.
• Apply protection controls to safeguard essential knowledge.
• Filter and sort collections to quickly find what matters to you.

This approach keeps knowledge organized without adding administrative overhead, making collections practical for everyday use rather than just large initiatives.

Designed for collaboration and sharing

Collections support collaboration by making it clear who can access and contribute to shared knowledge.

You can:

• Share collections with specific users or groups.
• Assign read or write access depending on the use case.
• Make collections available across the organization when appropriate.
• Integrate collections with supported messaging tools, such as Slack and Microsoft Teams, for team collaboration. 

Content is indexed while respecting your organization's existing security permissions, so a collection only ever surfaces what a given user is already authorized to see

These controls allow teams to collaborate confidently while maintaining clarity around ownership and permissions.

Powering AI-driven insights

Collections are the units of knowledge that the AI Assistant works against. By curating content into collections, you improve the relevance, accuracy, and usefulness of AI-generated responses.

Through collections, the AI Assistant can:

• Answer questions using only the documents and sources you select.
• Summarize key themes across large sets of content.
• Search for specific topics or details across multiple documents

Usage statistics help you understand which documents contribute the most value, giving visibility into how knowledge is actually used.

How collections fit into Knowledge Base

Collections sit at the center of the Knowledge Base experience:

• Content is added to collections.
• Permissions and sharing are managed at the collection level.
• AI interactions are scoped to collections.
• Reporting and usage insights are tied back to collections.

This makes collections the natural starting point for organizing knowledge and the primary hub for all downstream activity.

Related tasks and concepts

The following pages provide detailed guidance on working with collections:

• Overview of collections (this page)
• Create a manual collection
• Add content to a collection
• Edit a collection
• Share a collection
• Review documents in a collection
• Refresh a collection
• Re-index collection content
• Lock a collection
• Delete a collection
• Collection statistics