> ## Documentation Index
> Fetch the complete documentation index at: https://docs.parahelp.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Follow-up

> Your Customer Agent can send personalized follow-ups to customers.

Companies use follow-ups to let customers who requested a feature know it shipped, inform impacted users when a bug is fixed, and send updates during incidents.

The Internal Agent doesn't write each follow-up message. Instead, it gives your Customer Agent context, which then reads the ticket and writes a tailored reply to the customer.

<Danger>
  Follow-up actions are **sensitive**, so we recommend reviewing them before approval
</Danger>

## How you follow-up

<Steps>
  <Step title="Prompt the follow-up">
    Tell the Internal Agent what information you want to send and to whom.

    Based on your description, it will research and identify the relevant tickets to follow-up on and the instructions to give the Customer Agent so it replies appropriately with the right information.
  </Step>

  <Step title="Review the follow-up">
    The follow-up is staged as a single action in the "Actions" tab so that you can review it.
  </Step>

  <Step title="Following up">
    Once approved, your Customer Agent works on each ticket and sends a tailored follow-up to each customer.
  </Step>
</Steps>

## Follow-ups & Automations

It is common to run [Automations](/platform/automations) that identify relevant tickets to follow up on.

This could be whenever you merge a PR in GitHub. An automation kicks off and identifies whether the change is customer-facing. If so, it finds customers who requested the specific product update. Once you approve the follow-up action, your Customer Agent will let the customers know what shipped.

## Control

Personalized follow-ups need admin approval by default. An admin can enable auto-approval for a specific task or automation so they send without waiting. Because follow-ups can be sensitive actions, we recommend not enabling auto-approval unless you are confident in the task the Internal Agent performs.