Salesforce Agentforce Specialist Certification

Discover how the Einstein trust layer secures Salesforce data using generative AI, covering prompt and response journeys, data masking, zero data retention, response generation, and audit trails for LLM interactions.

Examine how the Einstein trust layer orchestrates prompt and response journeys, including dynamic grounding, data masking, prompt defense, toxicity detection, and audit trails between Salesforce apps and the language model.

Master secure data retrieval and dynamic grounding by fetching Salesforce CRM data from objects, fields, and related lists with permissions-based access to personalize prompts.

Mask sensitive data in large language model prompts by detecting patterns and context, replacing with placeholders, and preserving a secure mapping for later unmasking in AI responses.

Explore prompt injection and how prompt defense, data masking, and system policies shield generative AI applications, ensuring safe, reliable responses in the prompt journey.

Learn how Salesforce’s Einstein Trust Layer secures prompts through data grounding, masking, and defense, then uses a TLS-enabled gateway to connect Salesforce-created, external, or self-hosted models with zero data retention.

Explore response journey within the Einstein trust layer, where prompts are secured, toxicity detection scores content, data masking and unmasking use placeholders, and feedback audits stored in the data cloud.

Explore prompt builder concepts, including creating and using prompt templates—field creation, record summaries, sales email generation, and flex templates—plus grounding with CRM data, activation, execution, and required user roles.

Discover what a prompt is and what a reusable prompt template is, then learn how placeholders fetch CRM data to generate personalized prompts for AI models in Salesforce.

Explore the Salesforce prompt builder, a tool that creates ai prompts from CRM data. Build sales email, field generation, record summary, and flex templates with merge fields and generative ai.

Set up a developer account with Salesforce to practice the prompt builder, create and explore different types of templates, and learn the steps to verify email and log in.

Create personalized emails with a sales email template in Salesforce prompt builder, using merge fields from cases and contacts to quickly send case updates.

Use the record summary template to generate a concise overview of a case or other record by pulling essential fields such as case number, status, priority, and subject.

Activate the record summary template, fetch the case by id, run the case summary prompt, and display the response on the case page via a screen flow.

Explore temperature, frequency penalty, and presence penalty in prompt builder. Learn how temperature affects creativity, how frequency penalty reduces repetition, and how presence penalty encourages unique words.

Explore prompt template manager and prompt template user permission sets: manager creates, edits, and deletes templates; user uses existing templates; also overview Einstein sales email permission set.

Learn how to call a flow from a prompt template to fetch knowledge article data by issue type, then supply tailored resolutions to cases in an iPhone servicing context.

Learn to build a field generation prompt template in Salesforce that invokes a flow to fetch knowledge articles and auto-populate a case description with the recommended solution.

Explore how flex template in Salesforce Prompt Builder pulls data from up to five objects, related or not, to create newsletters and compare it with field generation and record summary.

Create two custom objects, features and recommendations, to demonstrate the flex template, defining fields like iPhone version (picklist) and iPhone features (rich text), then populate data.

Populate the features and recommendations objects in a flex template by creating and saving records for iPhone ten, iPhone 11, and iPhone 12. Build a newsletter from the populated data.

Create a flex template in Salesforce using prompt builder to define up to five sources and three objects—contact, features, and recommendations—then fetch fields to generate a personalized newsletter.

Create a flex template that pulls contact, feature, and recommendation data to generate and display a personalized newsletter within a screen flow on the contact page, based on iPhone versions.

Explore how Einstein Copilot, a natural language AI assistant for Salesforce CRM, uses topics and actions to fetch records, summarize data, draft emails, and troubleshoot with event logs.

Explore how Einstein Copilot, now called Agent Force, sits atop Salesforce CRM as an AI assistant with natural language processing, a reasoning engine, and a large language model.

Explore Einstein Copilot's natural language, conversational style, a library of standard and custom actions organized by topics, and planner service orchestration with Copilot builder and relevant permission sets.

Configure the agent in setup, create new topics, and route case-related queries to the topic and actions. Learn to summarize or retrieve case details, and escalate cases with Einstein Copilot.

Learn to create a flow type custom action to escalate a case, connect it to an agent action, and execute the flow at runtime to update the case status.

Learn to create a prompt template action and attach a custom action, enabling a Salesforce agent to fetch and summarize case records.

Create custom actions for the agent by building a flow that fetches a case by case number and returns fields. Use this to power a case summary action in conversations.

Explore retrieval augmented generation (rag), which combines retrieving relevant information and generating text, and learn how vectorization enables semantic search to deliver accurate, well-informed AI responses.

Explore how data lake objects store raw data and how data model objects organize it into DMOs, enabling a 360-degree view of customers and a search index with retriever.

Learn how Salesforce Data Cloud uses search index configuration and a retriever to fetch semantically related results by chunking data and storing vector embeddings.

Load contact data from Salesforce and Amazon S3 into Data Cloud, harmonize with matching rules, build a unified customer profile, and derive lifetime value insights for segmentation.

Create a Yahoo email to set up your Salesforce Trailhead account, then create your Salesforce org via the Trailhead link to enable Einstein and Data Cloud.

Create your Salesforce account using the provided org link, complete sign-up, verify email, reset password, and secure credentials to access the data cloud and Einstein setup.

Enable Einstein, create a connected app to obtain consumer key and secret, enable username-password OAuth with device flow, and configure scopes to connect Amazon S3 with Salesforce.

Create an Amazon S3 account and sign in to obtain the account ID, then build a data pipeline from S3 to Salesforce Data Cloud so file uploads trigger notifications.

Follow step-by-step setup for AWS S3 data cloud integration, including key pair, configuration file, shell script, and Lambda function, using a username-password flow and upcoming input parameter configuration.

Update the input parameters S3 configuration with your org username, login url, region, and unique bucket and folder names; the shell script builds the infrastructure and notifies Salesforce Data Cloud.

Upload the lambda function zip to cloud shell via s3 and configure the input parameters file with credentials. Update the lambda role, function name, and consumer key to deploy setup.

Obtain AWS credentials and run a shell script to deploy an S3 bucket, Lambda, and event notifications, guided by an input configuration file and CloudWatch logs.

Set up a data lake object from unstructured files in Amazon S3, establish a data cloud connection, and configure the unstructured data model to reflect uploaded files.

Connect Salesforce Data Cloud with Amazon S3 by testing data lake ingestion, configuring the instance URL, and validating updates in CloudWatch and the data lake object.

Upload files to amazon s3 data lake, explore with data explorer, and configure a search index and retriever for vectorized, semantic searches across data lake and data model objects.

Learn to create a search index and retriever for balance sheet data, manage ready status, and prepare data lake chunks and vector or hybrid indexes for the prompt template.

Learn to build a balance sheet prompt using the flex prompt template, free text sources, and a retriever in Einstein studio to answer questions from documents.

Master Einstein model builder to configure predictive and generative models, including custom and own large language models, through the prompt builder.

Learn to use Salesforce Einstein Model Builder in a no-code environment to train custom models with Salesforce data or external models, and contrast predictive models with foundational models like GPT-4.

Learn to create predictive and foundation models in model builder, connect external models (SageMaker, Vertex, OpenAI) or build from scratch.

Learn to build predictive sales models from scratch using diverse data, and deploy no-code AI with Einstein Copilot, Prompt Builder, and Model Builder in data cloud-enabled Salesforce.

Learn to use Einstein sales call summary and Call Explorer to summarize voice or video calls and to answer specific questions, with editable summaries including next steps and customer feedback.

Learn how Einstein tools in Service Cloud accelerate support: AI-generated chat replies with reply recommendation, email drafting from knowledge articles, and chat transcript summaries with work summaries.