Feature Overview
The Scheduled Task feature allows users to create periodic data analysis tasks using natural language. The system will automatically execute analysis and push results at the scheduled time. It can be used for daily anomaly detection, business data monitoring, trend insights, and other scenarios, helping users automatically discover abnormal changes in data and provide analysis suggestions.
Creating a Scheduled Task
Method 1: ASK AI Conversational Creation
| Step | Demo | Content |
|---|---|---|
| 1. User asks a question | ![]() | "Help me analyze whether yesterday's business data has any anomalies?" |
| 2. Auto-monitoring suggestion | ![]() | The system recognizes the user's anomaly monitoring needs and proactively recommends at the end of the analysis result, e.g.: "If such anomalies can be detected earlier, it can help the operations team intervene faster. I can help you set up daily automatic detection. Once a store rating falls below 3.0 or negative reviews exceed 20, I will notify you immediately and automatically analyze the cause." |
| 3. User confirms intent | ![]() | Reply "OK, set it up for me" |
| 4. Recommend monitoring metrics and configuration | ![]() | Based on the metrics and data distribution the user cares about, provide suggestions: monitoring metrics (rating, negative review count, order volume, revenue), thresholds, execution frequency (daily at 09:00), push strategy (only push on anomalies), and ask the user to confirm execution time, notification email, and monitoring metric scope |
| 5. User confirms configuration | ![]() | Confirm or adjust monitoring metrics, execution time, notification email, etc. |
Method 2: Manual Creation
-
Enter the "Scheduled Tasks" page
-
Click the "+ Scheduled Task" button in the upper right corner
-
Describe the task requirements through conversation; you can specify the email address directly
Result Notification
Email Push
- Enter the recipient email address in the task configuration
- After the task execution is complete, the system will automatically send the analysis results to the specified email address
Step 1: Email Notification Step 2: Key Results Check Step 3: View Full Analysis Process 


Manual Viewing
- Enter the "Scheduled Tasks" page to view the task list
- Click on a specific task to view historical execution records and analysis results

Task Details Page
- The task details page displays basic information and execution records:


Notes
-
Scheduled tasks are suitable for scenarios with high data update frequency (real-time or T+1). Scenarios with slower data updates (such as monthly reports) are not recommended for now.
-
Ensure the email address entered is correct; otherwise, you will not receive the push results.
-
Task execution results are automatically generated by the Agent. It is recommended to manually verify key conclusions.
-
You can enable or disable tasks in the task list at any time.





