Help Centre › Building Processes
This article describes the RPA Document Processing functionality to extract an organizational chart from a document or image.
Description
The Extract org chart function uses AI vision to interpret boxes, connector lines, and visual layout in an image, PDF, or PowerPoint file to determine the reporting hierarchy. It returns structured JSON with each person's name, title, and reporting relationship.
Supported file types
- Images: PNG, JPG, GIF, WebP
- PDF documents
- PowerPoint presentations (.pptx)
Setting it up
To use the Extract org chart function, follow these instructions:
{{popups:rpas/add:action=org-chart}}
Once on a process map:
- Navigate to the process where you would like to add the function
- Click on the green button's drop-down
- Click on "+ Activity"
- Select the DOCUMENT PROCESSING Extract org chart option
- Click "Add" once you are done
Input parameters
| Parameter | Required | Type | Default | Description |
|---|---|---|---|---|
| File path | Yes | Text | – | Path to the source file (image, PDF, or PowerPoint). |
| Page / slide | No | Number | 1 | Page or slide number for multi-page PDFs or PowerPoints. |
| Variable | No | Text | org_chart | State variable to store the result. |
Output variables
| Variable | Type | Description |
|---|---|---|
| {variable} | JSON | An object containing a nodes array. Each node has: id (string), name (person name), title (job title or null), and reports_to (parent node ID or null for top-level). |
Example output
{
"nodes": [
{"id": "1", "name": "Jane Smith", "title": "CEO", "reports_to": null},
{"id": "2", "name": "John Doe", "title": "CTO", "reports_to": "1"},
{"id": "3", "name": "Alice Brown", "title": "CFO", "reports_to": "1"}
]
}Errors
| Error | Description |
|---|---|
| File not found | The specified file does not exist. |
| Unsupported format | The file type is not supported. |
| Page out of range | The specified page or slide number exceeds the document length. |
| No chart detected | AI could not identify an organizational chart in the document. |
Tips
- Use high-resolution images for better accuracy.
- Ensure the org chart has clear boxes and connector lines for reliable hierarchy detection.
- For multi-page documents, specify the exact page containing the org chart.
- Access individual nodes in downstream steps using
{org_chart.nodes[0].name}.