AI-103 Implement Generative AI And Agentic Solutions Practice Question
You are implementing Retrieval-Augmented Generation (RAG) using Azure AI Search as the retriever and Azure OpenAI as the generator. Users report that the generated responses frequently miss critical context located in the middle of long retrieved documents. Which grounding optimization technique should you apply?
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
✓
Implement semantic ranking and reorder retrieved documents so the highest-scored chunks appear at the very beginning and end of the prompt context.
Lost-in-the-middle phenomena can be mitigated by placing the most relevant retrieved chunks at the beginning or end of the prompt context, or by employing ranker reordering strategies.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Increase the temperature parameter of the Azure OpenAI model to 1.2 to encourage creative synthesis of the middle content.
Why it's wrong here
Increasing temperature increases randomness and does not help the model attend to specific parts of the context window.
- ✗
Switch the Azure AI Search retriever from hybrid search to pure keyword BM25 search.
Why it's wrong here
Pure keyword search often performs worse than hybrid search combining vector embeddings and keyword matching.
- ✓
Implement semantic ranking and reorder retrieved documents so the highest-scored chunks appear at the very beginning and end of the prompt context.
Why this is correct
Language models tend to pay more attention to information at the beginning and end of a long context window.
- ✗
Reduce the chunk size to 64 tokens to force the retriever to return hundreds of tiny fragments.
Why it's wrong here
Extremely small chunk sizes remove necessary surrounding context and overwhelm the context window with fragmented data.
About these practice questions
This AI-103 question is part of Courseiva's 510-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed August 2026 · checked against the official Microsoft exam blueprint
This AI-103 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-103 exam.