Skip to main content
The End of Call Action runs a set of natural-language instructions automatically after every call ends, no matter how the call terminates. Unlike a webhook (which sends a single HTTP request), the End of Call Action runs as an autonomous step that can use your agent’s other functions — such as Send Email, Send SMS, Schedule, and any connected MCP tools — to complete what you describe. The agent is given the full transcript and the AI-generated summary of the call, then follows your instructions. Every action it takes is recorded in the call transcript so you have a complete record of what happened.
Each agent can have one End of Call Action. Its instructions can describe multiple steps (for example, “email a summary, then log the outcome to our CRM”). It only runs when the call produced a transcript, and it reuses your agent’s configured AI model.

​
Adding an End of Call Action

  1. Log into your account and navigate to the agent editor.
  2. In the Functions section, click Add and select End of Call Action.
  3. Enter your instructions, then save the agent.

​
Configuring the instructions

Describe what should happen after each call, and which of the agent’s functions to use. Because the action runs after the caller has hung up, it can only use functions that make sense post-call (email, SMS, calendar, MCP) — it cannot transfer, hang up, or take a payment.

​
Example Instructions

​
Use Cases

  • Post-call summaries: Email a summary of every call to a manager or shared inbox.
  • CRM logging: Record the outcome of each call in your CRM via an MCP integration.
  • Follow-ups: Automatically schedule a follow-up appointment or send a recap SMS.
If Secure Access is enabled for your organization, the End of Call Action still runs exactly as configured and may transmit call data (such as the transcript or summary) to external services through its functions (email, SMS, MCP). Only include instructions you intend to run automatically.