Power Platform Community Forum Thread Details

AI Builder's text recognition feature is powerful and easy to use, but currently it does not support recognizing text styles such as bold, italic, or colored text from images or documents.

This is a major limitation in real-world use cases like food labels, ingredient lists, or medical packaging, where bold text is used to highlight critical information (e.g., allergens, dosage, warnings).

Feature Request:

Please enhance AI Builder or Azure Document Intelligence to include text style metadata (e.g., bold, italic, font size, color) in OCR results.
For example:

{ "text": "Milk", "style": "bold" }

Use Cases:

Automatically extract allergen ingredients from food labels
Highlight important warnings or dosage in pharmaceutical documents
Improve document automation and classification based on visual emphasis

This feature would significantly improve the accuracy and usefulness of AI Builder in document automation scenarios.

Categories:

AI Builder

@지임-10051701-0

You're absolutely right in identifying a current limitation in AI Builder and Azure AI Document Intelligence: while both services offer powerful OCR capabilities for extracting text, tables, and key-value pairs, they do not currently support extracting text style metadata such as bold, italic, font size, or color from documents or images.

Current Capabilities

AI Builder is a low-code/no-code interface built on top of Azure Document Intelligence (formerly Form Recognizer)
2
.
Azure Document Intelligence supports:
- Text extraction (OCR)
- Key-value pair recognition
- Table extraction
- Custom model training
- Layout analysis (e.g., bounding boxes, reading order)