Build a Salesforce Agent to Chat with Unstructured PDF Data

Name: Build a Salesforce Agent to Chat with Unstructured PDF Data
Rating: 4.7 (158 reviews)

Master RAG, Einstein Data Library, Data Cloud, Prompt Builder, and AWS to Create Powerful Salesforce Agents

Highest Rated

Created bySalesforce troop, Amrita K

Last updated 11/2025

English

English [Auto],

What you'll learn

Understanding Einstein Data Library
Grounding AI Responses
How RAG Works
How Prompt builder and Agent works
Building and Deploying AI Agents
Real Time Projects
Working with Unstructured Data

Course content

4 sections • 35 lectures • 1h 51m total length

Introduction2:54
Disclaimer0:21
Salesforce Org Setup2:02
Important Note1:01
Enable Einstein1:48
Note: In recent Salesforce releases, some Setup menu names have changed. Follow these updated steps to enable the required features:
Open Setup
Click Setup. The Setup page opens in a new tab.
Enable Einstein
In the Setup Quick Find, search for and select Einstein Setup.
Click the Turn on Einstein toggle, and ensure Einstein is On.
Refresh your browser to reload Setup.
Enable Agentforce
In the Setup Quick Find, search for and select Agentforce Agents.
Toggle Agentforce at the top.
Below, enable Agentforce (Default) Agent as well.
Tip: If at any point you want to view a Topic or Action, note that these have now been moved under Agentforce Assets in Setup.
New way to enabling the Agents0:28
Data Cloud Setup2:12
Learn to enable data cloud in your Salesforce org by assigning the data cloud admin permission, refreshing, and running the data cloud setup to initialize the customer 360 data model.

AWS Signup2:43
Sign up for AWS and explore the free tier by creating a root user, verifying your email and identity via SMS, and providing billing information, then access AWS Management Console.
S3 bucket Setup1:53
Create an s3 bucket in aws, name it agent for this demo, then use the docs folder under agent force to illustrate folder structures and optional uploads.
Setup AWS Security (IAM)3:27
Upload Pdf To S30:57
Connect Data cloud to s3 bucket2:00
Quick Note --> How to view Connected App in New orgs0:13
Connected App3:42
How to Enable and Create a Connected App in New orgs
Go to Setup and type "External Client Apps" in the Quick Find box.

Click on External Client Apps → Settings.

Enable the checkbox: “Allow creation of connected apps”.

Click “New Connected App” to create your app.

Once created, your Connected App will also be visible under App Manager.
What is RAG and How it works?5:36
Explore retrieval augmented generation (RAG) and how it grounds prompts by ingesting unstructured data, chunking, vectorizing, indexing, and using a retriever to deliver up to date answers.
Configure unstructured data lake object4:36
Send File Notifications from Amazon S3 to Data Cloud3:53
Confirm Search index Status and DMOs Chunk3:25
If DMO chunk records are empty0:16
If Retriever is not created (Important Update)0:49
Create prompt Template12:11
Create a flex prompt template in the prompt builder to drive a RAG workflow using a retriever, Einstein Search, and the data cloud to answer user queries from PDFs.
Create an Agent and Agent Action9:40
Create a new Salesforce agent and LG support prompt template, configure LG FAQs, test the agent, and enable data cloud access for the Einstein service agent user.
Create a Digital Experience Site9:07
Route to Service Agent1:49
Embed Agent To the Site4:02
Embed the AI service agent on the site, publish the page, and test the chat widget. Customize the welcome and system messages, activate the agent, and verify the live interaction.
Delete your S3 Bucket1:26
Delete your S3 bucket by emptying it and permanently deleting it in the AWS console to prevent charges after finishing your project.

Demo1:29
Einstein Data Library and How it works2:51
Agentforce Data Library Setup3:09
Confirm Search index Status and DMOs Chunk1:03
Confirm the search index status and DMO chunks in the data cloud, verify the file DMO is ready with over 600 chunks, then pull data into the prompt template.
Important Update - Agentforce Data Library Custom Retriever0:50
Prompt Template5:44
Einstein Copilot Agent7:11
Publish Project 23:06
Publish project 2 by updating the Einstein Copilot system message, activating the Copilot, and testing within the Data Cloud app to reveal the Winter 25 release updates.

Requirements

Internet connection is needed to go through this course
Basic understanding of Salesforce is required before taking this course.
Trailhead Account

Description

Welcome to "Build a Salesforce Agent to Chat with Unstructured PDF Data", a project-driven course designed to equip you with the skills to create intelligent Salesforce agents that can effectively leverage unstructured data.

In this course, you’ll learn how to integrate technologies like AWS, Data Cloud, Einstein Data Library, and Prompt Builder to build agents capable of providing accurate, context-driven responses. We'll guide you through key concepts such as uploading data to AWS S3, injecting it into Salesforce Data Cloud, and structuring it into Data Lake and Data Model Objects for seamless data processing.

Using Prompt Builder, you’ll configure agents to fetch relevant information from Data Cloud based on user inputs and generate real-time, context-aware responses. You’ll also explore how the Einstein Data Library enables agents to process unstructured data for tasks like question-and-answer interactions.

This course emphasizes hands-on learning, with two exciting projects:

Project 1: Ingest unstructured data from AWS S3 into Salesforce Data Cloud, create a dynamic agent using this data, and expose it on an Experience Site for user access.
Project 2: Leverage Einstein Data Library to process and ground unstructured data, enabling agents to deliver precise and personalized responses.

You’ll gain practical experience integrating AWS S3 with Salesforce Data Cloud, managing Data Lake and Data Model Objects, and using advanced tools like Prompt Builder to enhance workflow automation.

By the end of the course, you’ll have the expertise to create intelligent agents that can transform unstructured data into actionable insights, improving efficiency, customer interactions, and overall productivity.

This course is perfect for developers, AI enthusiasts, and teams looking to integrate cloud solutions with Salesforce to tackle complex data challenges. Let’s dive in and unlock the potential of unstructured data!

Who this course is for:

Individuals Eager to Explore AI Agents
Developers Enhancing Expertise
Teams Managing Unstructured Data
Professionals Using Data Cloud Objects
For those eager to delve into AI agents and cloud technology to unlock new possibilities.
Ideal for developers aiming to upskill in handling unstructured data and automating processes efficiently.

Build a Salesforce Agent to Chat with Unstructured PDF Data

What you'll learn

Explore related topics

Course content

Introduction7 lectures • 11min

Project 1 - LG Support Agent - Unstructured Data from Amazon S319 lectures • 1hr 12min

Project 2: Agent for Release Notes with Einstein Data Library8 lectures • 25min

Retrieve Code in VS Code1 lecture • 3min

Requirements

Description

Who this course is for: