Fabrikam’s HR team wants to build an internal “search and ask” experience over policies, contracts, and FAQs. A solution architect says they should…

The statement is true. This architecture describes the Retrieval-Augmented Generation (RAG) pattern, which is a best practice for building enterprise AI solutions over private data. Using Azure AI Search to index internal content allows the system to first retrieve relevant, factual information from the company's knowledge base. This retrieved content is then passed to the Azure OpenAI model as context. This "grounds" the model's response in the company's specific data, significantly improving factual accuracy, reducing hallucinations, and enabling the system to cite its sources. It is superior to relying only on prompts with ad-hoc data chunks, which is less scalable and constrained by token limits.

False – would ignore the RAG best practice endorsed in Microsoft architecture guidance.

1. Microsoft Azure Architecture Center. (2024). Retrieval-Augmented Generation (RAG) in Azure AI Search. This document details the RAG pattern, stating, "RAG adds an information retrieval step to the generation process... grounding the model on external knowledge sources," validating the architect's statement.

2. Microsoft Azure Documentation. (2023). Azure Cognitive Search and LangChain: A Seamless Integration for Enhanced LLM Applications. Section: "Why RAG?". It explains that RAG is essential for "long-lived enterprise knowledge scenarios" because it provides the LLM with up-to-date, proprietary information that is not in its training data.

3. Walters, A., & Kukreja, P. (2023). Revolutionize your Enterprise Data with ChatGPT: Next-gen Apps with Azure OpenAI and Cognitive Search. Microsoft Build Session BRK212H. This session presents the RAG pattern with Azure AI Search and Azure OpenAI as the recommended architecture for enterprise search and ask scenarios.