Document Understanding Release Notes

DELIVERY:

Last updated Aug 26, 2025

January 2024

General Release Notes - Document Understanding^TM

26 January

Improvements

Previously, when classifying a document and clicking Next, the next document from the same category was displayed. This is now improved and the next unclassified document is displayed for a smoother classification process.

23 January

Active Learning in Document Understanding Public Preview

We are excited to announce that Document Understanding Active Learning is now in public preview. With this new interface and experience, model training is significantly improved by leveraging AI to guide you through the process with automatic annotation and expert-level recommendations on how to improve your model accuracy with the most informative datasets.

The new experience is organized into four components:

Build: Get your documents automatically classified and annotated, then follow the recommendations to achieve the highest model accuracy with active learning.
Measure: Measure your model performance and get further guidance on how to improve it with the most efficient model training.
Publish: Publish the projects and consume them in RPA or API automations.
Monitor: Monitor the performance of your automations and easily drill down into fine-grained document level audit.

Document Understanding API - Generative capabilities Public Preview

We are excited to announce that generative capabilities for Document Understanding API is now in public preview. You can use the following new capabilities:

Generative classifier to start classifying a document using a generative classifier.
Generative extractor to start extracting fields from a document using a generative extractor.

General Release Notes - Document Understanding^TM Classic Experience

23 January

Document Understanding API - Generative capabilities Public Preview

We are excited to announce that generative capabilities for Document Understanding API is now in public preview. You can use the following new capabilities:

Generative classifier to start classifying a document using a generative classifier.
Generative extractor to start extracting fields from a document using a generative extractor.