🧠 From Data Cloud to Agentforce: How to Use Enterprise Data in AI Workflows

by brandonin Advanced, Agentforce, Education, Hubspot, Learning, Management, Pardot, Uncategorizedon Posted on

Connecting the Dots: Empowering Teams with Data Cloud Libraries in Agentforce

Once you’ve connected data sources such as SharePoint into Salesforce Data Cloud, the next step is making that information actionable inside Agentforce. This guide walks you through the process of surfacing your data in Copilot conversations, workflows, and automations.

🔹 Organize Data with Libraries

Data Cloud Libraries are the foundation that Agentforce uses to access enterprise content.

  1. Navigate to Setup → Agentforce Data Libraries.
  2. Create a new Library or open an existing one.
  3. Add your data type (files, web or custom retriever).
  4. For custom connections with Data Cloud, use the Custom Retriever.
  5. Publish the library to make it available for AI.
Process Documents stored in Sharepoint containing Project process documents for ‘Agentforce Implementation – Sharepoint Integration’, ask Agentforce to: “Summarize the level of effort for implementing Sharepoint for Agentforce.”

🔹 How Agentforce Uses Data Cloud Libraries

When a user asks a question in Agentforce (via Copilot or a custom prompt):

  1. The Retrieval-Augmented Generation (RAG) looks at the connected libraries.
  2. Content is retrieved from the data source via the connector.
  3. Large documents are automatically chunked into smaller sections for semantic search.
  4. The AI composes an answer using the most relevant chunks.
Agentforce Prompt Builder Retrieval of Sharepoint Data. Result is PDF data being chunked and used in a query, then included in the AI response.

🔹 Use Data in Agentforce Prompts

With libraries connected, you can build prompts that directly tap into enterprise data. For example:

  • Process Documents stored in Sharepoint containing Project process documents for ‘Agentforce Implementation – Sharepoint Integration’, ask Agentforce to: “Summarize the level of effort for implementing Sharepoint for Agentforce.”
  • CSV Data stored in Amazon S3 containing product delivery & ERP Data, ask Agentforce: “What is the ETA for delivering X widgets to Georgia”

Agentforce can now blend structured CRM data and unstructured SharePoint content in a single response.

🔹 Automate with Agentforce Actions

Beyond Q&A, you can embed Data Cloud-powered insights into workflows:

  • Use CRM records from Data Cloud libraries to drive Next Best Actions: This is a core function of Agentforce and Next Best Action (NBA). Agentforce agents, using data from Data Cloud and the Einstein 1 Platform, can provide real-time recommendations to users based on customer data. For example, when a service agent views a customer’s record, Agentforce could recommend an upsell opportunity or a retention offer based on that customer’s history.
  • Enable copilots for service teams that pull knowledge base + case data in real time: Agentforce is the underlying “agentic” AI platform for Salesforce’s Copilot. For service teams, Copilot can use a Retrieval Augmented Generation (RAG) framework to pull from the company’s knowledge base, case history, and other relevant data sources in real time. This information is then used to generate content, summarize records, and help resolve customer issues

🔹 Final Thoughts

By connecting data sources into Data Cloud and then exposing them through Libraries, you unlock their full potential in Agentforce. This architecture ensures:

  • ✅ Secure, governed access to enterprise data.
  • ✅ Fast retrieval through chunking and semantic search.
  • ✅ Seamless blending of structured + unstructured content.

With this setup, Agentforce stops being a general AI assistant and becomes a context-aware enterprise copilot — one that can answer, summarize, and act using your company’s own data.

CONNECT WITH CYPRESS

Looking to Implement Agentforce for your Organization? Setup some time with our Experts to start planning your next Agentforce project