You are here:
How the Create Agent from Bot Tool Works
With the Create Agent From Bot tool, the move to Agentforce is quick and easy. After creating an AI agent from an Einstein bot, the bot stays active so that you can make the move at your own pace.
Required Editions
| Available in: Lightning Experience. |
| Available in: Enterprise, Performance, Unlimited, and Developer Editions with the Einstein for Service, Einstein Platform, or Agentforce for Service add-on. |
| To use enhanced Messaging channels with AI agents, see required editions. |
The Create Agent From Bot tool walks you through three stages.
- Checking that required permissions and licenses are in place before the Create Agent from Bot tool launches.
- Gathering details about your bot’s elements, including its dialogs, action steps, and intents.
- Previewing and selecting specific dialogs to create agent topics and actions from, with a summary of each dialog.
- Generating agent topics and actions based on the bot’s details.
- Defining basic agent settings such as the name, role, and company. The agent’s behavior is determined by the topics and actions generated from the selected dialogs.
All you do is follow along. After the tool creates the agent, you can customize, test, and activate the agent.
Some bot elements aren’t moved to the agent created by the Create Agent from Bot tool. See Considerations for Creating Agents from Bots.
How Agent Topics Are Generated
After gathering details about your bot, the tool identifies which dialogs or dialogs groups it can create topics from. Typically, the tool uses all dialogs that have intents and utterances. Then, Agentforce groups similar intents and dialogs together and, based on each group, generates a classification description, scope, and topic instructions. If a dialog has custom conditional logic, that logic is included in the topic instructions. Learn more about topics.
How Agent Actions Are Generated
After identifying dialogs to create topics from, the tool creates agent actions based on action steps in those dialogs. For each agent action, Agentforce generates a description and input and output parameters. Learn more about actions.
- Action steps that only help the bot navigate conversations
- Action steps that create feedback logs
Bot Dialogs with Flow-Based Action Steps
- Service Request dialog with an action step that creates a service request
- Get Service Status dialog with an actions step that retrieves a service request
- Reschedule Service dialog with an action step that updates a service request
- Customer Details dialog
- Vehicle Details dialog
Agent Topic with Agent Actions
Service Appointment Management topic with a Manage Service Appointment, Get Service Request, and Get Car Models action.
