You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
description: Discover how Foundry IQ provides a managed knowledge platform built on Azure AI Search that eliminates the need to build custom RAG infrastructure for every agent.
description: Learn how to configure different data sources for Foundry IQ knowledge bases, including Azure AI Search, Blob Storage, Web, SharePoint, and OneLake options.
description: Learn how to configure agent instructions to control retrieval behavior, test different query types, and ensure consistent citation and grounding in production.
description: Test your knowledge of Foundry IQ and knowledge-enhanced AI agents.
7
+
ms.date: 01/15/2026
8
+
author: madiepev
9
+
ms.author: madiepev
10
+
ms.topic: unit
11
+
module_assessment: true
12
+
durationInMinutes: 3
13
+
content: |
14
+
quiz:
15
+
questions:
16
+
- content: "What is the primary advantage of Retrieval Augmented Generation (RAG) over simple AI agents?"
17
+
choices:
18
+
- content: "RAG eliminates the need for large language models by relying entirely on document retrieval."
19
+
isCorrect: false
20
+
explanation: "Incorrect. RAG augments large language models with retrieval capabilities—it doesn't replace them. The agent still uses its language model to generate responses, but now grounds them in retrieved organizational content."
21
+
- content: "RAG enables agents to ground responses in current organizational information and provide source transparency."
22
+
isCorrect: true
23
+
explanation: "Correct. RAG delivers three critical advantages: real-time updates that keep agents current, source transparency that shows which documents informed each response, and factual grounding that eliminates fabricated information."
24
+
- content: "RAG automatically retrains the language model whenever organizational documents change."
25
+
isCorrect: false
26
+
explanation: "Incorrect. RAG doesn't retrain the model. Instead, it retrieves current information at query time and augments the agent's context, allowing it to stay current without retraining."
27
+
- content: "Which data source option provides real-time access to SharePoint content with Microsoft 365 governance?"
28
+
choices:
29
+
- content: "SharePoint Indexed, which pre-processes SharePoint content into Azure AI Search."
30
+
isCorrect: false
31
+
explanation: "Incorrect. SharePoint Indexed creates a pre-processed index for faster search and advanced features, but it's not real-time. The indexed content depends on the indexing schedule."
32
+
- content: "SharePoint Remote, which queries SharePoint sites and libraries in real-time."
33
+
isCorrect: true
34
+
explanation: "Correct. SharePoint Remote provides real-time queries to SharePoint without pre-indexing, respects existing SharePoint permissions, and always accesses current content."
35
+
- content: "Azure Blob Storage, which connects to SharePoint files stored as blobs."
36
+
isCorrect: false
37
+
explanation: "Incorrect. Azure Blob Storage is a separate data source for files stored in Azure Storage containers, not for accessing SharePoint content."
38
+
- content: "What is the purpose of scoring profiles in Foundry IQ knowledge bases?"
39
+
choices:
40
+
- content: "To encrypt sensitive fields and protect confidential information during retrieval."
41
+
isCorrect: false
42
+
explanation: "Incorrect. Scoring profiles influence search relevance, not data security. They boost specific fields or attributes so more important results surface first during retrieval."
43
+
- content: "To boost specific fields or attributes so more important results surface first."
44
+
isCorrect: true
45
+
explanation: "Correct. Scoring profiles let you define field weights (like boosting title matches 3x over content matches) and freshness functions (like prioritizing documents modified within the last 90 days) to influence ranking."
46
+
- content: "To configure how documents are chunked and embedded for semantic search."
47
+
isCorrect: false
48
+
explanation: "Incorrect. Chunking and embedding happen during indexing when you add data sources. Scoring profiles affect retrieval ranking after content is already indexed."
49
+
- content: "Why is it critical to specify retrieval behavior in agent instructions?"
50
+
choices:
51
+
- content: "Without proper instructions, agents might answer from training data instead of the knowledge base, provide unverifiable responses, or fail to cite sources."
52
+
isCorrect: true
53
+
explanation: "Correct. Effective instructions specify when to retrieve (always use the knowledge base), how to cite (exact format for source attribution), and what to do when unsure (fallback behavior when information isn't found)."
54
+
- content: "Instructions determine the semantic ranking algorithm that Foundry IQ applies to search results."
55
+
isCorrect: false
56
+
explanation: "Incorrect. Semantic ranking is configured at the knowledge base level, not in agent instructions. Instructions control agent behavior—how it decides to search, cite, and respond."
57
+
- content: "Instructions enable the agent to automatically update knowledge base content when it detects outdated information."
58
+
isCorrect: false
59
+
explanation: "Incorrect. Agents consume knowledge from knowledge bases but don't modify them. Knowledge bases update when you change the connected data sources, not through agent instructions."
title: Exercise - Integrate an AI agent with Foundry IQ
5
+
description: Hands-on exercise to create a Foundry project, set up a knowledge base, and integrate an AI agent with Foundry IQ for knowledge-enhanced interactions.
6
+
ms.date: 01/15/2026
7
+
author: madiepev
8
+
ms.author: madiepev
9
+
ms.topic: unit
10
+
title: Exercise - Integrate an AI agent with Foundry IQ
0 commit comments