Update Custom Document Intelligence Model for New Format

Implement an Azure Document Intelligence in Foundry Tools solution
Answer Correct answer: C — Add the additional contract format to the existing training set and retrain the model.

You have an app named App1 that uses a custom Azure AI Document Intelligence model to recognize contract documents. You need to ensure that the model supports an additional contract format. The solution must minimize development effort. What should you do?

  1. Lower the confidence score threshold of App1.
  2. Create a new training set and add the additional contract format to the new training set. Create and train a new custom model.
  3. Add the additional contract format to the existing training set. Retrain the model. Correct Answer
  4. Lower the accuracy threshold of App1.

Community Votes

C
100%

100% of anonymous learners picked answer C. Votes are pick records left by other test-takers — they are not the verified answer.

Community Insight

Adding new document examples to an existing training set and retraining minimizes development effort while expanding model recognition capabilities.

To support an additional contract format in an Azure AI Document Intelligence custom model, you must add the new format to the existing training set and retrain. This page establishes that retraining an existing model minimizes development effort compared to creating a new one.

Choosing to create a new training set and model (Option B), which unnecessarily duplicates effort and discards existing trained capabilities.

Community Discussion (5 comments)

syupwsh 👍 1 Selected: C
Option C directly addresses the need to support an additional contract format without starting from scratch. By adding examples of the new contract format to the existing training set and retraining the model, you enhance the model's ability to recognize the new format while leveraging the already trained capabilities for existing formats. This minimizes development effort compared to creating and training a completely new model. C it is
reigenchimpo 👍 2 Selected: C
It is obvious that C is the correct solution.
reiwanotora 👍 1 Selected: C
It must be C.
Murtuza 👍 3 Selected: C
Given answer C is correct
Murtuza 👍 1
Add the additional contract format to the existing training set and retrain the model: This is a more efficient solution. By augmenting the existing training data with the new format, you can fine-tune the model without starting from scratch. Retraining the model with the updated dataset should help it adapt to the additional format

Comments & Corrections

No comments yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Submitted for moderation before publishing. Keep it helpful and respectful.

Expert Analysis

Why the Answer Is Correct

Adding the additional contract format to the existing training set and retraining the model (Option C) directly teaches the current model to recognize the new format. This approach leverages the model's existing knowledge of the original contract formats, requiring only the incremental effort of labeling new examples and retraining. It perfectly satisfies the requirement to minimize development effort.

Why the Other Options Are Wrong

Option A and Option D suggest lowering thresholds (confidence or accuracy), which only changes the application's filtering logic but does not teach the underlying model how to understand the new contract format. Option B suggests creating a completely new training set and model, which significantly increases development effort by forcing you to start from scratch rather than building upon existing work.

Community Comment Notes

The community strongly agrees that updating the existing training set is the most efficient path. As syupwsh noted, "Option C directly addresses the need to support an additional contract format without starting from scratch." Murtuza also highlighted that "augmenting the existing training data with the new format, you can fine-tune the model without starting from scratch."

Official Reference

Exam Strategy

When asked to minimize development effort for expanding an existing custom model's capabilities, always prefer updating and retraining the existing model over creating a new one. Avoid threshold adjustments when the core issue is a lack of model knowledge.

Related Analysis

← Back to AI-102 Study Guide