Brief about Agentforce

Agenforce is a trusted conversational AI assistant built into the Salesforce interface. Agentforce is Salesforce’s generative AI-powered assistant designed to enhance productivity and customer interactions. It helps users perform tasks efficiently within Salesforce by leveraging Topics and Actions:

Also note that Agentforce (Default) is the new name for the agent type, previously known as Einstein Copilot for Salesforce.

To set up the Agents Naviagte to Setup – Agent Studio –> Agents

Language settings within Copilot Builder allow you to change the tone of your messages.

There are three options:

Formal: polished and professional

Neutral: unbiased and nonevaluative

Casual: relaxed and informal

Agent – Out-of-the-box Actions

• Answer Questions with Knowledge – “What is the policy for returns over 30 days?”
• Create Close Plan
• Draft or Revise Sales Email – “Help me write an intro email to Steve from Acme.”
• Explore Conversation – “What was the customer sentiment on the call?”
• Find Similar Opportunities
• Get Forecast Guidance
• Identify Object by Name – “List the opportunities for the Acme account”
• Identify Record by Name – ““Show me the Acme records”
• Query Records (Beta) – “Find all open opportunities set to close this quarter sorted by created date.”
• Query Records with Aggregate (Beta) – “How many opportunities were created in the past 5 days?”
• Send Meeting Request
• Summarize Record – “Create a summary for the Acme deal.”
  

If none of the standard actions fulfill your requirements, we can create custom actions via Copilot Builder. These custom actions can be built using Apex actions, flows and API support which is really powerful.

Einstein Copilot is Salesforce’s generative AI-powered assistant designed to enhance productivity and customer interactions. It helps users perform tasks efficiently within Salesforce by leveraging Topics and Actions:

• Topics: Represent the primary subject or intent of a user’s request. Examples include “Case Management,” “Lead Generation,” or “Order Tracking.” Topics help categorize interactions to ensure relevant responses.
• Actions: Define specific operations Einstein Copilot can perform within a given topic, such as “Create a Case,” “Update an Opportunity,” or “Schedule a Meeting.” Actions allow users to complete tasks seamlessly using AI-driven automation.

By integrating Topics and Actions, Einstein Copilot streamlines workflows, reduces manual effort, and enhances customer engagement through natural language interactions. The below image shows some of the standard topics – Account management, General FAQ and Manage Deals.

Topic – Details

The below topic details image shows the topic configuration and other details of the topic. Such as

Scope – thats tells the agent-copilot what can be done in the current scope of that topic in Natural language

Instructions – helps copilot to take decisions how to to use topics for different use cases. Also, help placing guardrails.

Topic Actions: This displays all the list of Actions -copilot able to do as part of the current Topic.

Knowledge : This is another important tool through copilot refer the data library , knowledge articles to respond to user questions.

Language: On this tab – all the copilot supported languages can be selected along with the Tone selection

Event Logs; All the copilot interaction log can be viewed on the event logs view.

Conversation Preview:

The preview feature shows the topics picked by the copilot to respond to users question and the corresponding actions executed with in the defined scope

along with the instructions defined.

This preview feature also helps to validate the capabilities and inefficiencies of the Copilot before releasing . It helps to find out the gaps and adjust as per the use cases by modifying the scope, adding instructions and also adding custom topics and actions .