DP-800: Developing AI-Enabled Database Solutions

By Microsoft · Question Mode

QuestionQ32

Implement AI capabilities in database solutions

You have an Azure SQL database containing a table named knowledge_base. knowledge_base stores human resources (HR) policy documents and has columns named title, content, category, and embedding.

You have an application named App1. App1 queries two relational tables, employee_profiles and benefits_enrollment, that contain HR data. App1 hosts a chatbot that calls a large language model (LLM) directly.

Users report that the chatbot correctly answers general HR questions, but gives outdated or incorrect answers when policies change. It also cannot answer questions that refer to internal policy documents by title or category.

You need to recommend a Retrieval Augmented Generation (RAG) solution that resolves the chatbot issues.

What should you recommend?

Select
Retrieve grounding data from:
Inference step to perform the retrieval:
Explanation

The policy-document table is the grounding source because it contains the current policy content and its embeddings. A RAG inference flow embeds the user query with the embedding model and uses vector similarity search to retrieve the most semantically relevant policy documents; that retrieved content is supplied to the LLM as grounding data. This avoids relying on stale model training and supports semantic matches beyond exact keywords.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!