A legal firm wants to build a RAG solution that answers questions based on their internal case documents. They have the documents in Azure Blob Storage. They plan to use a prompt flow in Microsoft Foundry. Which components must they include in the flow?

An index lookup tool to retrieve relevant document chunks, and an LLM call that uses the retrieved chunks as context.

A content safety tool and an LLM call, but no retrieval step.

A prompt template that contains all the case documents as static text.

An LLM call only, with the entire document set included in the prompt.

A travel agency is building a prompt flow in Microsoft Foundry to recommend travel itineraries. The flow currently uses a single LLM call. They want to add a step that checks the user's budget and filters recommendations accordingly. They also want to use a prompt template to structure the LLM call. What should they do?

Create a new prompt flow for each budget range.

Use a content safety tool to filter recommendations based on budget.

Modify the LLM call to include the budget in the system prompt, but do not add any filtering step.

Add a Python tool to the flow that filters the LLM output based on the budget, and use a prompt template with a budget input.

A government agency is planning a generative AI solution in Microsoft Foundry. They have strict data residency requirements that mandate all data processing and storage remain within a specific geographic region. What should they do during the planning phase?

Select a Foundry hub region that matches the required data residency location.

Store all data on-premises and only use Foundry for model inference.

Use a model that is available globally to avoid regional restrictions.

Defer the decision until after deployment.

A manufacturing company wants to build a RAG solution that answers questions from their equipment manuals. The manuals contain many tables and diagrams. The company wants to minimize the number of tokens sent to the model while still retrieving relevant content. What should they do?

Skip embedding and use keyword search only.

Chunk the documents by fixed character count and embed each chunk.

Use a layout-aware chunking strategy that preserves tables and diagrams, then embed the chunks.

Embed the entire manual as a single vector.

A financial services company has built a prompt flow that generates investment summaries. Before deploying to production, they need to ensure the flow does not produce hallucinated facts about specific stocks. Which evaluation approach should they use?

Check the token usage of the flow to estimate cost.

Manually review a few sample outputs from the flow.

Measure the latency of the flow to ensure it is fast enough.

Use a groundedness evaluation metric that compares the output to the source data.

Microsoft Certified:Azure AI Engineer Associate

Domain 2Objective 1

Build Generative AI Solutions with Microsoft Foundry AI-102 Practice Questions (Page 4)

Part of the Implement generative AI solutions domain, which accounts for 15-20% of the AI-102 exam. Microsoft does not publish an official question count, but from its 100-minute exam (~40–65 total, ~6–13 in this domain), expect 2–4 from this objective — we provide 45 practice questions to prepare you well beyond it. (estimate)

45questions here

9free pages

8concepts

15-20%of the exam

Questions 16–20

16application · medium
A legal firm wants to build a RAG solution that answers questions based on their internal case documents. They have the documents in Azure Blob Storage. They plan to use a prompt flow in Microsoft Foundry. Which components must they include in the flow?
Select an answer first
17application · medium
A travel agency is building a prompt flow in Microsoft Foundry to recommend travel itineraries. The flow currently uses a single LLM call. They want to add a step that checks the user's budget and filters recommendations accordingly. They also want to use a prompt template to structure the LLM call. What should they do?
Select an answer first
18application · medium
A government agency is planning a generative AI solution in Microsoft Foundry. They have strict data residency requirements that mandate all data processing and storage remain within a specific geographic region. What should they do during the planning phase?
Select an answer first
19application · medium
A manufacturing company wants to build a RAG solution that answers questions from their equipment manuals. The manuals contain many tables and diagrams. The company wants to minimize the number of tokens sent to the model while still retrieving relevant content. What should they do?
Select an answer first
20application · medium
A financial services company has built a prompt flow that generates investment summaries. Before deploying to production, they need to ensure the flow does not produce hallucinated facts about specific stocks. Which evaluation approach should they use?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to AI-102

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Microsoft. “AI-102” is a trademark of its owner, used for identification only.