Data Library for Agentforce
Data Library lets Agentforce admins turn enterprise documents into searchable knowledge bases that ground AI agent responses. We built it to solve a trust problem: without visibility into how data becomes agent knowledge, admins couldn't diagnose bad answers or improve accuracy.
Discover data sources
The Data Library home page is the starting point for grounding agents in enterprise knowledge. It surfaces every supported data source — uploaded files, Salesforce Knowledge, external MCPs, Google Drive, SharePoint, Slack — so admins can see what's available, what's already connected, and what they can add. No hunting through separate tools; discovery and connection happen in one place.
Data Processing
Data pipelines are a black box. Users upload files, wait, and hope. The status card makes the process visible: it surfaces each step — upload, index, retriever setup — in real time, so users can track progress, spot failures early, and know the moment their knowledge base is ready for agents. The status framework allows any data source pipeline to plug into it so users gets consistent experience across all data source type.
Context block in Agent Builder
When users connect data inside the Agent Builder, they care about speed — not pipeline transparency. The context block lets them quickly attach data sources for processing without leaving the agent building flow.
Impact
2.9 → 6.2
User confidence after redesign (out of 7)
40%
Reduction in customer support cases — admins self-serve
~80%
Increase in product adoption