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

# Automation Overview

> Automate AI work with scheduled tasks

Cloosphere's automation runs repetitive AI tasks at scheduled times without human intervention. Use cron-based scheduled tasks to automate report generation, data analysis, status monitoring, and more — and receive results via email or webhooks.

<Frame caption="Schedule list — card-based schedule view">
  <img src="https://mintcdn.com/cloocus/aj5f2DPkGcweQdgX/images/en/automation/schedules-list.png?fit=max&auto=format&n=aj5f2DPkGcweQdgX&q=85&s=eb9b858b3b2fe969d31a3edbef318502" alt="Schedule list — card-based schedule view" width="2000" height="935" data-path="images/en/automation/schedules-list.png" />
</Frame>

***

## Automation Features

<Columns cols={2}>
  <Card title="Schedules" icon="clock" href="/en/automation/schedules">
    Auto-run agents, flows, and models on a cron-based scheduler. Receive results via email, Slack, Teams, etc.
  </Card>

  <Card title="Future Plans" icon="lightbulb">
    Event-trigger-based automation, conditional workflows, and other features are planned.
  </Card>
</Columns>

***

## Automation Scenarios

| Scenario                        | Target           | Frequency      | Notification        |
| ------------------------------- | ---------------- | -------------- | ------------------- |
| **Daily sales report**          | Sales DB agent   | Daily 9 AM     | Email               |
| **Weekly data analysis**        | Analytics agent  | Every Monday   | Slack webhook       |
| **System status check**         | Monitoring agent | Hourly         | Teams on failure    |
| **Monthly KPI report**          | KPI flow         | First of month | Email + Slack       |
| **Real-time anomaly detection** | Anomaly agent    | Every 10 min   | Webhook (immediate) |

***

## Automation Flow

```mermaid theme={null}
flowchart LR
    subgraph Configuration
        A[Pick target] --> B[Write prompt]
        B --> C[Set frequency]
        C --> D[Connect notification channel]
    end

    subgraph Execution
        E[Cron trigger] --> F[Run agent/model]
        F --> G[Generate result]
    end

    subgraph Delivery
        G --> H[Save to chat]
        G --> I[Send email]
        G --> J[Send webhook]
    end

    D --> E
```

***

## Key Concepts

### Target

The resource to auto-run. Pick one of agent, flow, model, or dashboard.

| Target Type   | Description                           | Best For                              |
| ------------- | ------------------------------------- | ------------------------------------- |
| **Agent**     | AI with Knowledge Base/DB connections | Data analysis, document-based reports |
| **Flow**      | Multi-step workflow                   | Complex multi-step automation         |
| **Model**     | Base LLM                              | Simple text generation, summarization |
| **Dashboard** | Dashboard HTML export                 | Periodic dashboard snapshots/reports  |

### Cron Expression

Standard format defining the run frequency. `0 9 * * 1-5` means "weekdays at 9 AM". An intuitive cron editor is provided — you don't need to write it manually.

<Tip>
  Cron expressions run in the **Timezone** field shown in the editor, not UTC. This defaults to your browser's local timezone, so you enter your local wall-clock time directly — no UTC offset conversion needed.
</Tip>

### Delivery Channel

Channels to receive run results. Supports email, Slack, Teams, Discord, and more.

***

## Prerequisites

<Warning>
  To use scheduled tasks, the following must be in place:

  * Admin must enable `features.scheduled_tasks` permission for the user's group
  * To create schedules, admin must set the group's `workspace.schedules` level to **write** (default is read — view only)
  * You must have access permission to the agent/flow/model being auto-run
  * For email/webhook notifications, admin must pre-configure channels in **Admin > Settings > Channels**
</Warning>

***

## Get Started

<Steps>
  <Step title="Prepare workspace">
    First prepare the [agent](/en/workspace/agents), [flow](/en/workspace/flows), or model to auto-run. For data-analysis automation, create an agent connected to a [database](/en/workspace/database).
  </Step>

  <Step title="Create a schedule">
    On the [Schedules](/en/automation/schedules) page, set target, prompt, run frequency, and notification channel.
  </Step>

  <Step title="Test run">
    Use the **Run Now** button to test settings. If the result isn't what you expected, edit the prompt and retest.
  </Step>

  <Step title="Operate and manage">
    A schedule becomes **active as soon as it's created** and runs automatically per the configured frequency. If the test results look good, leave it as is; to pause it or change settings, deactivate it from the more menu (⋮) and then edit.
  </Step>
</Steps>

***

## Example Setup: Daily Sales Report Automation

```mermaid theme={null}
flowchart LR
    A[Sales DB Agent] --> B[Run daily at 9 AM]
    B --> C[Query sales data + generate chart]
    C --> D[Save result to chat]
    C --> E[Email to executives]
    C --> F[Notify Slack #sales channel]
```

| Setting            | Value                                                                     |
| ------------------ | ------------------------------------------------------------------------- |
| **Target**         | Sales analytics agent (DB-connected)                                      |
| **Prompt**         | "Analyze today's sales data and write a report including DoD growth rate" |
| **Frequency**      | Daily 9 AM (`0 9 * * *`)                                                  |
| **Notification 1** | Email (on success) → executives                                           |
| **Notification 2** | Slack webhook (always) → #sales                                           |

<Tip>
  Charts generated by DbSphere agents are server-side rendered to images and auto-included in emails and webhooks.
</Tip>