Multi-Surface Example: Ground an Agent with an Agentforce Data Library

Complete all the prerequisites in Headless Examples.

In your terminal, run sf org list. If you see the headless org alias in the list, you’re successfully authenticated.

This example also requires:

• Data 360

Copy and paste this prompt into your AI assistant:

1Create a Q&A agent called "QnA".

Your AI assistant creates the agent. We’re not ready to publish and activate it yet.

Agentforce Data Libraries (ADL) enhance the accuracy of AI features, such as Agentforce agents, by connecting them to your trusted data sources. Data libraries use unstructured or semi-structured data. They convert large sets of information from the web, documents, or text fields into useful, searchable information.

Download these files to your filesystem so that you can use them to ground the agent. The files contain information about a fictitious DiamondVision TV from Acme Corp.

1. order-tracking.txt
2. product-troubleshooting.txt
3. technical-support.txt

Create a file-based data libary Agentforce data library.

Copy and paste this prompt into your AI assistant:

1Create an Agentforce data library. Add the order-tracking.txt, product-troubleshooting.txt, and technical-support.txt files to the data library.

Your AI assistant creates the data library with the three sample files.

Wire the new Agentforce data library into the QnA agent.

Copy and paste this prompt into your AI assistant:

1Ground the QnA agent with the new Agentforce data library.

Your AI assistant connects the data library to the agent to improve the accuracy of agent responses.

Copy and paste this prompt into your AI assistant:

1Test the agent with realistic customer scenarios and interactions.

Your AI assistant generates a test plan and executes the tests. If any tests fail, review the root causes returned by your AI assistant and ask it to apply the recommended fixes.

Copy and paste this prompt into your AI assistant:

1Commit and activate the agent in the org.

Your AI assistant publishes and activates the agent.

Complete steps 1 and 2 in Headless Examples.

This example also requires:

• Data 360

Complete these steps by opening Salesforce org in your browser.

1. From the App Launcher, enter Agent, and then select Agentforce Studio.
2. On the Agents tab, click New Agent.
3. Select an Agentforce Service agent template. Under Agent Details, select or create an Agent’s User Record for this agent. See Configure Service Agent Access for details.
4. Enter QnA as the agent’s name and then click Let’s Go.

Agentforce Data Libraries (ADL) enhance the accuracy of AI features, such as Agentforce agents, by connecting them to your trusted data sources. Data libraries use unstructured or semi-structured data. They convert large sets of information from the web, documents, or text fields into useful, searchable information.

Download these files to your filesystem so that you can use them to ground the agent. The files contain information about a fictitious DiamondVision TV from Acme Corp.

1. order-tracking.txt
2. product-troubleshooting.txt
3. technical-support.txt

Create a file-based Agentforce data library.

1. From Setup, in the Quick Find box, enter data library, and then select Agentforce Data Library.
2. Click Add Data, then Upload Files.
3. Name the library File Data Library, and the API name is automatically populated. Optionally, enter a description.
4. Click Upload Files and add the three files downloaded in the previous step.
5. Click Save and wait for the Status of each file to say Uploaded.

Wire the new Agentforce data library into the QnA agent.

1. Return to Agentforce Studio and open the QnA agent.
2. In the Explorer tab, scroll down to Data and select Data Library.
3. Select File Data Library from the dropdown.
4. Save your agent.

1. Click Preview.
2. Select your test mode.
3. To start a preview conversation, in the chat window, enter a typical user question or request. See the example interaction below.
4. To review the agent’s behavior and reasoning, start with the Interaction Summary in the Summary tab. The Interaction Summary shows you a high-level overview of the steps your agent took to generate a response, including the result of each. Hover over each step for more details.
5. To get AI assistance, click Ask Agentforce in the Interaction Summary panel. We send the integrated Agentforce assistant information about the preview session so you can ask it questions.
6. If the test reveals errors or gaps, ask the assistant to make updates, or make changes yourself in the Canvas view. If your agent version is committed, make a new version first and modify it.
7. To test changes, refresh the session by clicking the reset icon on the Preview panel. Continue testing, modifying, and retesting until you’re satisfied with the agent’s behavior.

1. Click Commit Version, then click Commit again.
2. Click Activate, then click Activate again.