Overview
Configure your agents to access AI Data Hubs so they can retrieve organizational knowledge and provide contextually accurate responses. This tutorial shows how to connect agents to document collections and knowledge bases within your workspace.
Learning Objectives:
Navigate to agent configuration settings
Select specific data hub collections for agent access
Publish agent changes to activate data hub connections
Verify agents can leverage data hub content in responses
Prerequisites
You'll Need:
Access to a workspace on Starter, Best-of-Breed, Professional, or Enterprise plan
At least one configured agent
Existing AI Data Hub collections with uploaded documents
Publisher or administrator role for the agent you want to configure
Configuring Agent Data Hub Access
Grant Data Hub Access to Your Agent
Navigate to the agents section located in the left panel
Click the agent name you want to configure to open its settings page
Scroll to the "Data Hub Access" section at the bottom of the agent settings
Select the collections and data hubs you want the agent to access
Check the box next to each relevant collection name
Use the search field to filter collections if you have many
Click the "Publish" button in the top-right corner of the agent settings page
Result: The agent updates immediately and can now search the selected data hubs when responding to queries.
Note: The data hub selection interface matches the one on the chat screen, providing a consistent experience across the workspace.
Verify Agent Access and Test
Return to the home page by clicking the workspace logo or "Home" in the top navigation
Select your configured agent from the right panel sidebar
Ask a question related to content in the connected data hubs
Review the agent's response for data hub citations or source references
Result: The agent's response includes information from your data hubs, confirming successful integration.
Best Practices
Select only relevant collections: Limit agent access to data hubs directly related to their assigned tasks. Focused knowledge improves response accuracy and prevents confusion.
Review access quarterly: Audit which data hubs your agents can access every 90 days. Update selections to reflect new collections and remove outdated ones.
Create agent-specific collections: For sensitive or confidential information, create dedicated data hub collections instead of granting agents access to your entire library.
Test agent responses after changes: After publishing data hub updates, test the agent with sample queries to verify it retrieves information correctly.
Document your configuration: Maintain a simple log of which agents access which data hubs for easier troubleshooting and team knowledge sharing.
Common Questions
Q: Can agents access data hubs in real-time as content updates?
A: Yes. Agents search data hubs dynamically with each query, so newly added documents become available immediately after upload and indexing.
Q: How many data hubs can I connect to a single agent?
A: There is no technical limit, but best practice recommends connecting 3-7 relevant collections to maintain response quality and speed.
Q: Will agents reveal sensitive information from private data hubs?
A: Agents only access data hubs you explicitly select. Create separate collections for sensitive content and limit agent access to those collections carefully.
Q: Can I see which data hub an agent used to answer a specific question?
A: Many implementations include source citations or references in agent responses.
Q: Do I need to republish the agent after adding new documents to a connected data hub?
A: No. New documents in existing collections are automatically indexed and available to agents without republishing.
Q: What happens if I remove a data hub from an agent's access?
A: The agent immediately loses access to that collection's content. Existing conversations won't be affected, but future queries won't retrieve information from the removed data hub.
Q: Can different agents access different data hubs?
A: Yes. Each agent's data hub access is configured independently, allowing you to tailor knowledge access based on each agent's specialized purpose.