AI-102 Plan and manage an Azure AI solution Practice Question
Your company is developing an AI-powered document processing solution using Azure AI Document Intelligence. The solution must extract data from scanned PDF forms. The forms are in a custom format not supported by prebuilt models. You have 10,000 labeled forms for training. The solution must be deployed in a region that supports Document Intelligence and must be accessible via a REST API. You need to ensure the solution can process forms with high accuracy. What should you do?
⚠ Common exam trap
Many candidates confuse Azure AI Language's entity extraction with Document Intelligence's form extraction, or assume that a prebuilt model can be adapted via manual mapping, when in reality custom training is mandatory for unsupported formats.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Train a custom extraction model using the labeled forms
Azure AI Document Intelligence supports training custom extraction models using labeled forms, which is essential for handling custom form layouts not covered by prebuilt models. With 10,000 labeled forms, you have sufficient data to train a high-accuracy model that extracts specific fields via the REST API, meeting the deployment and accessibility requirements.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Use Azure AI Language to extract entities from the text
Why it's wrong here
Azure AI Language extracts entities from existing text; it does not perform OCR on scanned PDFs, so no text reaches the model. It is tempting because entity extraction handles unstructured content well, and would be correct once documents are already digitised as text.
- ✓
Train a custom extraction model using the labeled forms
Why this is correct
Custom extraction models learn the layout and field patterns of your specific form type from labelled samples, which prebuilt models cannot handle. Training with 10,000 labelled forms yields the high accuracy the custom format demands, and the model is callable via REST API.
- ✗
Use a prebuilt model and map fields manually
Why it's wrong here
Prebuilt models target standard document types, so their schemas cannot represent the custom form's fields regardless of manual mapping. It is tempting because prebuilt models need no training data, and would be correct if the forms matched a supported type such as invoices or receipts.
- ✗
Use the Read model and write custom logic to extract fields
Why it's wrong here
The Read model performs OCR and returns text lines without field extraction, so custom logic cannot match a trained custom model's accuracy on 10,000 labelled forms. It is tempting because Read is cheap and needs no training, and would suit scenarios where only raw text is required.
Go deeper
Related to this question
About these practice questions
Courseiva writes every AI-102 question from scratch — 761 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AI-102 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-102 exam.