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

# Scheduled tasks

> A scheduled task runs an agent's analysis automatically on a schedule and pushes the conclusion to Feishu or email, ideal for periodic monitoring and reports. This page covers configuration, run history, and how it differs from a channel.

## What is a scheduled task

A scheduled task makes an [agent](/en/concepts/agent/overview) **run an analysis automatically on a schedule** and push the conclusion to you. You decide which analysis to run, how often, and where the result goes — and it runs itself, with no one watching.

## Typical scenarios

* Every day at 9 AM, analyze yesterday's key order metrics and push them to a Feishu group
* Every Monday, summarize last week's sales and send it to a mailbox
* At the start of each month, generate a monthly business report

It's a good fit for periodic monitoring and reports.

## Configuration

| Field            | Description                                                                                                                   |
| ---------------- | ----------------------------------------------------------------------------------------------------------------------------- |
| **Instruction**  | The analysis for the agent, in natural language, e.g., "Analyze yesterday's orders and output the key metrics"                |
| **Frequency**    | Set with a cron expression, e.g., daily `0 9 * * *` or every Monday `0 9 * * 1`. **The interval must be at least 10 minutes** |
| **Notification** | Where results are delivered: a Feishu webhook or email                                                                        |
| **Enabled**      | A toggle; you must set a notification method before enabling                                                                  |

Each scheduled task belongs to an agent, which runs the analysis.

## Running and history

* **Trigger now** — Run once without waiting for the schedule, to verify the configuration
* **Run history** — Every run is logged, with its conclusion, status (success / failure), and time
* **Next trigger** — Shows the next automatic run time

## Scheduled tasks vs channels

Both connect to an agent, but in different directions:

* **Channel**: the user starts it — a two-way conversation, you ask and it answers
* **Scheduled task**: the system starts it on a schedule — a one-way push, it runs automatically and sends you the result

## Frequently asked questions

**Q: Where do results go?**
A: A Feishu webhook (group bot) or email — pick one.

**Q: What's the shortest interval?**
A: 10 minutes minimum.

**Q: What happens if a task fails?**
A: The run history records the failure status and reason (analysis failed / delivery failed) so you can troubleshoot.