The AI Specialist Certification is tailored for professionals who want to leverage the capabilities of generative AI within the Salesforce platform. Those in this role are key to implementing and managing Einstein AI solutions, optimizing business processes with the help of advanced tools like Copilot Builder, Prompt Builder, and Model Builder.

This certification is particularly suited for Salesforce Administrators, Developers, and Architects, who utilize both standard AI functionalities and custom enhancements to drive innovation within their organizations.

There are 5 areas of knowledge that are covered by the Salesforce AI Specialist certification.

AI Specialist Topic Weighting Chart

The following are the core topic areas of the  AI Specialist certification and what you’re expected to know:

This topic includes the following objectives:

       The Einstein Trust Layer is a secure AI architecture that is designed to enhance the security, privacy, and trustworthiness of generative AI features in Salesforce. It incorporates Secure Data Retrieval & Dynamic Grounding to ensure that prompts for AI outputs are based on dynamically retrieved CRM data that the executing user can access. Data Masking is utilized to protect sensitive information in prompts by replacing it with placeholder text during the prompt journey. Prompt Defense guards against unintended responses and potential attacks.
The Zero Data Retention policy guarantees that data used by external partner model providers is deleted after processing. After a response is generated by an LLM, Toxicity Detection is used to scan it for harmful or inappropriate content. Data Demasking is used to de-mask the sensitive data in the generated response. A user can provide their feedback on the response. Audit Trail containing audit and feedback data are stored in Data Cloud.

      The Einstein Trust Layer can be implemented and managed by configuring several key features within Salesforce. Einstein Generative AI is enabled on the Einstein Setup page and serves as the foundation for the Trust Layer. Data Masking is automatically enabled to safeguard sensitive information, though administrators can adjust the data masking configuration of specific sensitive data attributes. Audit Trail can be enabled by turning on the collection and storage of generative AI audit data on the Einstein Feedback page in Setup.  Additionally, various pre-built Data Cloud reports and dashboards allow users to analyze audit and feedback data and track key metrics like data masking and content toxicity. Automated notifications can be set up for specific types of audit and feedback data by creating record-triggered flows. By configuring these features, administrators can maintain compliance, security, and privacy while utilizing Einstein generative AI capabilities.

This topic includes the following objectives:

          Einstein Generative AI features for sales offer powerful capabilities to boost sales teams' productivity and effectiveness. Using Einstein Copilot, an AI-powered assistant, sales reps can generate record summaries, product recommendations, close plans, etc. Sales Emails allows users to send personalized, AI-generated emails to leads and contacts. Call Explorer enables users to extract critical insights from voice and video calls, such as competitor mentions and product inquiries. Call Summaries allows users to automatically generate detailed post-call summaries, which include next steps and feedback. Other key generative AI features include Einstein Coach, Sales Signals, and Automatic Contact Enhancement. Additionally, other Sales Cloud Einstein features, such as Opportunity Scoring, Lead Scoring, and Einstein Forecasting, help optimize sales strategies. The combination of these features empowers sales teams to work smarter, save time, and improve the overall sales performance.

         Various Einstein Generative AI features can be utilized in customer service scenarios. Einstein Service Replies for Email and Einstein Service Replies for Chat can be implemented when agents need AI-generated responses to handle customer inquiries efficiently. Service AI Grounding ensures that the responses are grounded in case context or the company’s knowledge base. Einstein Work Summaries can be set up to predict and fill case summaries, issues, and resolutions. Einstein Reply Recommendations is the ideal choice for providing relevant responses to agents in chat and messaging sessions based on the org’s closed chat transcripts. For scenarios involving the creation of new cases, Einstein Case Classification helps by automatically predicting key fields such as Priority, Reason, and Type. Other generative AI features for service include Einstein Article Recommendations, Einstein Bots, Case Wrap-Up, Knowledge Creation, Conversation Mining, and Einstein Next Best Action.

This topic includes the following objectives:

            Prompt Builder is a tool that allows users to create detailed prompt templates that instruct large language models (LLMs) to produce AI-generated responses. A prompt template consists of a prompt that utilizes various elements like ingredients, guidelines, and grounding.
Prompt Builder offers four types of reusable prompt templates—Sales Email, Field Generation, Record Summary, and Flex—that can be configured to meet specific business needs. The Prompt Template Workspace allows users to draft and test these prompt templates.
The Draft with Einstein button allows users to draft emails using a sales email prompt template. Users can click an icon next to a dynamic form field to use a field generation prompt template and populate the field with AI-generated text. In addition, invocable actions and flows provide seamless integration of prompt templates with workflows. 

           Salesforce users can be allowed to manage and execute prompt templates by assigning the appropriate permission sets. The Prompt Template User permission set can be assigned to users who need to access and run prompt templates. It provides visibility to the generative AI-enabled icon next to a dynamic form field that uses a field generation prompt template.
The Prompt Template Manager permission set can be assigned to users who need to create and manage prompt templates. The Einstein Sales Emails permission set can be assigned to those who need to draft emails using a Sales Email prompt template. Additionally, it is necessary to ensure that users can access all the relevant fields when drafting sales emails. 

         When creating a prompt template, it's important to consider best practices, limitations, and other considerations. A well-crafted prompt template should be clear, concise, and consistent. Best practices include roleplaying as a character to provide context and including instructions surrounded with triple quotes (""") in a separate section. There are certain limitations to be aware of. The User and Organization related lists cannot be used as merge fields, and the Activities related list is not supported for objects like Account and Case. Certain numerical limits, such as the maximum number of related list merge fields, also apply to prompt templates. Furthermore, a Lightning record page must be upgraded to Dynamic Forms to use field generation prompt templates, and data grounding must be accurate and complete for the expected results.

        When configuring prompt templates in Salesforce, selecting the appropriate grounding technique is essential to ensure accurate and relevant AI-generated content. Grounding techniques connect a prompt template to data sources such as Salesforce records, related lists, Apex classes, and external data. For instance, a record merge field links a template directly to an object field, while a related list merge field is used to ground a template using an object's related list.
A flow merge field can introduce dynamic logic by triggering a flow for a more complex use case. A record snapshot can be used to ground a prompt with data available on the user's page layout for an object. Prompt templates can also utilize Apex merge fields for programmatic use cases and DMO merge fields to use data stored in Data Cloud. Moreover, Retrieval Augmented Generation (RAG) can be used to ground templates with unstructured data, such as knowledge articles, emails, and chat transcripts. 

        Using Prompt Builder effectively requires understanding how to create, activate, and execute prompt templates. When creating a new prompt template, an AI specialist can specify its type, name, and description. Depending on the selected type, additional fields may be required. The Prompt Template Workspace provides a space to write, preview the resolution and response, and test a prompt. It also allows users to select the model configuration. An edited prompt template can be saved as a new version or new template, but it must be activated to make it available to users. The execution of a prompt templates depends on its type. A sales email prompt template can be run by clicking the Draft with Einstein button. A field generation prompt template can be run by clicking a generative AI-enabled field icon. A flex prompt template can be run using an invocable action, Connect REST API, or Connect in Apex. Consistently generating effective prompts is an iterative process, requiring best practices like testing multiple responses.

This topic includes the following objectives:

       Einstein Copilot is a conversational AI assistant built seamlessly into the Salesforce user interface that automates tasks and answers questions by executing predefined actions based on user requests. It is particularly beneficial in scenarios where businesses need to streamline operations, improve efficiency, and provide quick and accurate responses to users. The core components of Einstein Copilot are topics, actions, the reasoning engine, and the copilot itself.
By implementing Einstein Copilot, organizations can automate common tasks, make data-driven decisions, and provide a consistent, efficient CRM experience for Salesforce users. Its ability to execute standard and custom actions makes it adaptable for various business needs. Einstein Copilot can be activated using the Agent Builder and offers flexibility for different business requirements.

           The large language model (LLM) plays a critical role in executing Einstein Copilot actions based on user requests. The LLM identifies the user's intent, determines the best matching actions in the right order, and generates an appropriate response while maintaining the conversation flow. The reasoning engine is responsible for orchestrating this process by launching topics and actions, ensuring the right tasks are performed to meet the user's request. Additionally, session event logs in the Agent Builder can be accessed to debug and analyze the sessions associated with the execution of copilot actions. Einstein Copilot supports OpenAI GPT-4o for planner service calls. Copilot actions support making calls to other predefined LLMs.

          Einstein Copilot can leverage standard and custom actions to execute tasks on behalf of Salesforce users. Standard actions, such as Find Similar Opportunities and Summarize Record, are included with Einstein Copilot by default and help users complete common tasks. Custom actions, on the other hand, offer flexibility to tailor the copilot’s capabilities to specific business needs. When creating a custom action, an AI specialist can select a Reference Action Type, which can be Apex, Flow, or Prompt Template. Copilot actions can be assigned to topics in Agent Builder. Instructions can be specified to define how a custom action should be used during conversations. Properly managing and assigning these actions ensures that Einstein Copilot delivers the most relevant and effective responses to user requests.

           To effectively manage and monitor the adoption of Einstein Copilot, it's important to track user interactions, feedback, and usage analysis. Agentforce Analytics provides a dashboard and reports that offer insights into Copilot's usage, feedback, and adoption, helping refine its effectiveness. The User Input Analysis dashboard provides insights into how users engage with Copilot, their requests, and whether the AI assistant is meeting them. Event Logs in Agent Builder allow an AI specialist to track detailed logs of user interactions to optimize user experiences and resolve common copilot issues.

This topic includes the following objectives:

      Model Builder is a powerful tool for creating and connecting AI models to address specific predictive and generative AI requirements. It supports creating predictive AI models that support both binary classification and regression use cases, allowing organizations to leverage historical data for predictive insights. It also enables integration with LLMs hosted on external AI services, such as those hosted on platforms like Amazon and Google Cloud. Furthermore, the Models API facilitates access to large language models (LLMs) from key partners, empowering teams to harness AI capabilities such as generating chat conversations and text responses using Apex methods or the REST API.

        Salesforce provides several options to configure large language models (LLMs) for generative AI use cases. An AI specialist can leverage Salesforce-enabled foundation models, such as OpenAI GPT-4, which come pre-configured and can be easily used with Prompt Builder. The Bring Your Own Large Language Model (BYOLLM) functionality allows adding new foundation models for unique business requirements. BYOLLM requires configuration details such as the model's URL and authentication information. Model configurations based on foundation models can be created or edited. Temperature, Frequency Penalty, and Presence Penalty, known as hyperparameters, can be adjusted to fine-tune a model configuration.