Get started in the next knowledge area of the exam as we prepare to learn Agentforce Concepts

In this lesson, we tour the Agent Builder and begin to get familiar with the various components which make up an Agent, such as Topics and Actions. We preview a conversation to see the inner working and decision making of an agent and test it with various scenarios.

The first learning objective found in the exam guide for the second knowledge area states that you need to be able to explain how an agent works in Agentforce. I share a helpful few paragraphs inside of the New Agent screen that spells out how an agent works, using Topics and Actions.

We create an Agent for Setup Agentforce Agent and test it out.

Here I share a few other types of Agents that are available in demo orgs for Salesforce partners and may eventually become available in the free dev orgs.

Here we generate test data to test an Agent for Setup Agent. I also leverage a template to do further testing in addition to the generated test cases.

The default Agentforce Setup Agent fails to retrieve permission set assignments for a user, establishing the need for our custom solution, which I run through in this lesson.

Custom Topic Instruction Text:

1. When an admin asks which permission sets are assigned to a specific user mentioned by name (like "Mike Wheeler"), ALWAYS use the IdentifyRecordByName action FIRST to get the user's Salesforce ID.

2. For the IdentifyRecordByName action, use "User" as the object type and the person's name as the search term.

3. After successfully identifying the user with IdentifyRecordByName, use the output User ID as the input for the Get_User_Permission_Sets action.

4. Present the permission set information in a clear, organized format, listing each permission set name and type (Standard or Custom).

5. If the user does not have any permission sets assigned, clearly inform the admin and suggest checking profile permissions instead.

6. If multiple users are found by the IdentifyRecordByName action, ask the admin to clarify which specific user they want information about.

7. If the admin directly provides a User ID or username, you may skip the IdentifyRecordByName action and use Get_User_Permission_Sets directly.

8. For security audits or when the admin requests details about specific permission sets, provide additional context about what types of access these permission sets typically provide.

You are unable to access sandboxes or change sets in the free developer orgs. So in this lesson, I demonstrate the process of creating and uploading a change set which includes agents, prompt templates and many more generative AI components.

In this lesson, we look at various Agentforce related Reports and Dashboards which help you to manage and monitor agent adoption in your org.

Create a custom Record Summary Prompt Template and put it to the test. Learn how to edit, version and activate prompt templates.

We draft a Sales email for a lead record and review the other types of Prompt Templates you'll need to understand on the exam.

We create a field generation template which will generate descriptions on account records.

In this lesson, we review the Prompt Template User and Prompt Template Manager Permission Sets

In this lesson, we update our page layout to a dynamic form and associate our Field Generation Prompt Template with the Description field on the Account page layout. We then put our Prompt Template to the test and further iterate and refine it based on the results we receive.

In this lesson, we create a new user account and grant them Lightning Experience User rights through a Profile Adjustment. We also enable allowing admins to log in as other users in our org as well as turning on the Enhanced Profile User Interface via User Management Settings in Setup. We then log in as our new user without the Prompt Template User Permission Set Assigned and we then verify that our generative AI enabled field for Account Description does not display the Spark (AI) icon. We then assign the Prompt Template User Permission Set to our new user account and upon logging in as them can now see the Spark icon and execute prompt templates as that user.

In this lesson, we sign up for an Einstein AI enabled Trailhead Playground org which has Sales Email Templates and Einstein for Sales enabled. This functionality is locked down in our free dev orgs. Please note that these Trailhead Playground orgs only last for a few days and then they expire, so be sure to only sign up for one when you are ready to put it to use and get hands on to gain this specific experience. With our new playground, we enable Einstein via Einstein Setup as well as Einstein for Sales. This gives us a lot of Standard Sales Email Templates, which I copy from in order to create a new custom Sales Email Template. We then activate it and put it to use in the interface, also being sure to select a newer model to use with our new custom Sales Email Template.

In this lesson, we create a Flex template, using an example template from Salesforce. Once we put it to the test, we encounter an issue, which requires further refinement and iteration and testing of the prompt until we receive the desired results.

In this lesson, we explore various techniques and best practices to get better results out of our prompts. We test the placement of instructions, the use of various models, as well as turning masking off on certain data points. We also refer to a recent change of Salesforce's where they have Data Masking now turned off and they provide this reasoning for that change - Data masking through the Einstein Trust Layer is disabled to improve the performance and accuracy of agents. All data accessed by agents, including personally identifiable information (PII), is protected in transit and isn't stored or used for training purposes by external LLM providers, as part of our strict zero-data retention policy.

In this lesson, we create an Agentforce Data Library for Standard Operating Procedures and upload a Flow Screen SOP to it.

In this lesson, we set up Knowledge and update the Knowledge Article page layout. We create a custom field for the article body content as a rich text area field and create a knowledge article.

Welcome to this course. In this course you will learn all of the key topics you need to understand in order to pass the Salesforce Certified AI Associate Certification.

In this lesson, we explore the Salesforce Certified AI Specialist Exam Guide as well as the Trailmix and other training resources provided by Salesforce. We will leverage these resources to extract topics using Generative AI to gain our scope and focus for our studies.

In this lesson, I share a link to where you can access a Free Data Cloud org with Salesforce Einstein enabled. NOTE: These free orgs only provide you with 14 days of access before they expire, so work through this course within 2 weeks!

In this lesson, we begin the setup of the Data Cloud and turn on Einstein Generative AI in our org.

In this lesson, we review the Einstein Trust Layer and the journey of a prompt in Salesforce Einstein. I share the Einstein Trust Layer diagram as a resource.

We also configure the Data Masking for the Einstein Trust Layer in our org to protect sensitive data.

In this lesson, we enable Einstein Feedback in our org to collect and store Einstein Generative AI Audit Data.

In this lesson, we enable Einstein for Sales and review several of the core Sales features that leverage Generative AI in Einstein. We also review permissions related to Einstein for Sales.

In this lesson we enable Einstein for Service and set up Prompt Grounding.

In this lesson, we turn on the Einstein Data Library.

In this lesson, I verify that we cannot enable Knowledge in our free Einstein Learning orgs, therefore we cannot ground prompts on Knowledge articles.

In this lesson, we generate a Sales email using Einstein Copilot leveraging a standard Sales Email Prompt Template. We then inspect that template in the Prompt Builder.

In this lesson, we create a Custom Sales Email Prompt Template. We review the resulting Resolution and Response and review Masking Details by using sample data.

In this lesson, we further revise our Custom Sales Email Prompt Template to use the Open AI ChatGPT 4o model and further revise our prompts to provide greater flexibility to the end user and not require them to replace placeholder text with replacement text.

In this lesson we activate and test our Custom Prompt Template.

In this lesson, we create a Field-Generation Template Triggered Prompt Flow.

In this lesson we create a Field Generation Prompt Template.

In this lesson I get you familiar with the recent renaming of the Copilot Studio to the Agent Studio and Agentforce.

In this lesson we create a custom Copilot Action. I also share a best practices resource for you.

To use an action with a copilot, add it to a topic. An action can be added to multiple topics.

In this lesson, I compare and contrast Global Actions vs. Agent Actions to clear up potential confusion.

In this lesson, I capture the instructions available in our Topic for later reference.