Amazon
Anthropic
Avaya
Cisco
CompTIA
CrowdStrike
EC-Council
EXIN
Fortinet
Google
ISC
ISC2
ITIL
Juniper
Microsoft
NVIDIA
Palo Alto Networks
Salesforce
SAP
VMware
All Vendors

Microsoft Developing AI Apps and Agents on Azure AI-103 Dumps in PDF

Free Microsoft AI-103 Real Questions (page: 4)

AI-103 PDF — 187 Questions

Case Study
This is a case study. Case studies are not timed separately from other exam sections. You can use as much exam time as you would like to complete each case study. However, there might be additional case studies or other exam sections. Manage your time to ensure that you can complete all the exam sections in the time provided. Pay attention to the Exam Progress at the top of the screen so you have sufficient time to complete any exam sections that follow this case study.
To answer the case study questions, you will bed to reference information that is provided in the case. Case studies and associated questions might contain exhibits or other resources that provide more information about the scenario described in the case. Information provided in an individual question does not apply to the other questions in the case study.
A Review Screen will appear at the end of this case study. From the Review Screen, you can review and change your answers before you move to the next exam section. After you leave this case study, you will NOT be able to return to it.
To start the case study
To display the first question in this case study, select the “Next” button. To the left of the question, a menu provides links to information such as business requirements, the existing environment, and problem statements. Please read through all this information before answering any questions.
When you are ready to answer a question, select the “Question” button to return to the question.
Overview
Company Information
Contoso, Ltd is a multinational retail company that builds, deploys, and manages generative AI and agent-based solutions by using Microsoft Foundry.
Existing Environment
Identity Environment
Contoso uses Microsoft Entra ID for identity management, authentication, and authorization capabilities that enable agents to access organizational resources and services.
Contoso recently formed a new AI engineering team named Agent1Dev Team to optimize and maintain existing AI solutions.
The team collaborates with solution architects, DevOps engineers, and security engineers to design,
implement. monitor, and secure AI applications.
Contoso also has a team named Agent1Test Team that is responsible for validating AI solutions before the solution deployments.
Generative Environment
Contoso has a Microsoft Foundry deployment that contains two projects named Project1 and Project2.
Project1
Project1 contains a customer support agent named Agent1 that assists customers with product inquiries and troubleshooting requests.
Agent1 has the following configurations:
▪ Agent1 uses a base model deployment.
▪ A safety evaluation pipeline is NOT enabled.
▪ Tool invocation approval workflows are NOT enabled.
▪ Conversation memory constraints are NOT configured.
Agent1 interacts with customers by using digital support channels and answers general questions about
Contoso products.
Project1 is deployed to an Azure region located in the European Union (EU).
Agent1Dev Team will use Project1 to optimize and maintain Agent1.
Project2
Project2 contains a deployed video generation model. The marketing department at Contoso has access to
Project2 and plans to use the model to develop a video creation solution.
Development of the solution is incomplete.
Data Environment
Contoso stores product-related information in Azure resources that support AI applications.
The Azure environment contains an Azure Blob Storage account named storage1 that stores product detail sheets for all the Contoso products.
The product sheets include specifications, feature descriptions, and product support information that Agent1
can use to answer customer questions. The product sheets are stored in the PDF format.
Problem Statements
Contoso identifies the following issues:
-Agent1 has only general knowledge of the Contoso products.
-A recent chat interaction with Agent1 was analyzed for sentiment. The results of the analysis have NOT
-been processed yet.
-Agent1 does NOT use the detailed product information in the product sheets stored in storage1 when responding to customer questions.
-The finance department at Contoso reports that vendor invoices must be reviewed manually to ensure that the invoices match the terms defined in the vendor contracts. The invoices contain tables, logos, and varied layouts that make the documents difficult to process consistently.
Requirements
Planned Changes
Contoso plans to implement the following changes:
-▪ Implement a solution for Project1 that analyzes the vendor invoices by evaluating both the visual layout and the textual content of the invoices, so that the invoice details can be verified against the vendor contract terms.
-▪ Update the base model deployment used by Agent1 and standardize the model version to ensure continuity and consistent responses.
▪ Enable Agent1 to retrieve and use the detailed product information from the product sheets stored in storage1.
▪ Implement an indexing solution for the product sheets that Agent1 can use to answer customer questions.
▪ Complete the development of the video creation solution.
Technical Requirements
Contoso identifies the following technical requirements:
-▪ The model deployment used by Agent1 must support scalable, high-throughput generative AI workloads and dynamically scale to handle variable customer support traffic, without requiring reserved throughput capacity.
-▪ The product sheets must be processed by using an indexing pipeline that enables semantic and vector search, so that Agent1 can retrieve the relevant product information.
-Responses generated by using the product sheet information must be relevant, complete, and accurate.
-Agent1 must be able to use the product sheets to answer natural language questions about product details.
-The model version used by Agent1 must remain consistent to ensure stable responses.
-The data processed by the model must remain within the EU.
Security and Compliance Requirements
Contoso identifies the following security and compliance requirements:
-▪ API keys must NOT be used to access Foundry-deployed models.
-▪ Access to the Azure resources must follow the principle of least privilege.
-▪ The developers at Contoso must authenticate to Microsoft Foundry resources by using Microsoft Entra authentication.
-▪ Access to Project1 must be assigned to the members of Agent1Dev Team by using a security group named
-SC_Agent1_Dev.
-▪ Access to Project1 must be assigned to the members of Agent1Test Team by using a security group named
-SC_Agent1_Test.
-▪ Agent1 must never reveal customer information, even if a document that contains customer data is added erroneously to the product sheet repository in storage1.
-The product sheets might contain images that include embedded text. Agent1 must be protected from malicious instructions potentially hidden within the images.
Business Requirements
Contoso identifies the following business requirements:
-Users that interact with Agent1 must have a personalized experience in future interactions, including the ability for Agent1 to retain conversation context and recall relevant information from previous interactions.
-Agent1 must answer questions only about the products sold by Contoso.

Note: This section contains one or more sets of questions with the same scenario and problem. Each question presents a unique solution to the problem. You must determine whether the solution meets the stated goals. More than one solution in the set might solve the problem. It is also possible that none of the solutions in the set solve the problem.
After you answer a question in this section, you will NOT be able to return. As a result, these questions do not appear on the Review Screen.
You have a multimodal AI generative model that accepts image uploads and uses extracted image text to generate responses.
You discover that users can upload unsafe images and embed hidden instructions into images to manipulate the model.
You need to implement controls to mitigate the risk.
Solution: You configure protected material detection.
Does this meet the goal?

  1. Yes
  2. No

Answer(s): B

Explanation:

Correct:
* You configure a prompt shield for documents.
Prompt Shield for Documents: Highly Effective (Critical Defense)
How it helps: This shield specifically scans untrusted, third-party data inputs (like external documents or text extracted from uploaded images).
Mechanism: It evaluates the extracted image text before it is sent to the LLM to identify hidden jail
* You configure a prompt shield for user prompts.
Prompt Shield for User Prompts: Partially Effective (Defense in Depth)
How it helps: This shield targets direct jailbreak attempts written manually by the user in the text prompt field accompanying the upload.
Mechanism: It prevents the user from typing supporting instructions that prime the model to execute the hidden instructions found within the image.
* You configure image moderation to block unsafe content before processing the images.
Implementing rigorous image moderation is one of the most effective ways to secure multimodal AI systems against these threats. Moderation acts as a necessary gatekeeper, preventing malicious inputs from ever reaching the generative model.
Incorrect:
* You configure protected material detection.
Protected Material Detection: Ineffective for this Threat
Why it does not help: This feature is designed to scan model outputs to prevent the generation of copyrighted text, proprietary source code, or licensed imagery.
Limitation: It does not scan inputs for adversarial instructions and will not prevent a user from manipulating the model's logic.


Reference:

https://www.upgrad.com/blog/what-is-multimodal-ai/
https://learn.microsoft.com/en-us/azure/ai-services/content-safety/concepts/jailbreak-detection



You have a Microsoft Foundry project.
You plan to build a customer support solution that contains an agent. The solution must meet the following requirements:
-▪ Provide accurate, context-aware responses grounded in internal product documentation stored in Azure AI
Search.
▪ Require deep, multi-step reasoning across long contexts.
▪ Generate detailed natural language responses.
Which type of model should you use to power the agent?

  1. a multimodal model
  2. a small language model (SLM)
  3. a key phrase extraction model
  4. a large language model (LLM)

Answer(s): D

Explanation:

For a support agent requiring deep multi-step reasoning, long context processing, and strict grounding in internal documents, the GPT-5.5 (or GPT-5.5-Pro) model from the Azure AI Foundry model catalog is the best choice.
Here is how GPT-5.5 directly addresses the requirements:
Deep Multi-Step Reasoning: These frontier models use deliberate reasoning and iterative planning before generating a response. This allows the agent to decompose complex support queries, analyze alternatives,
and prevent hallucinations without requiring heavy custom prompt engineering.
Grounded, Accurate Responses:
Rather than doing this alone, pair the model with Foundry IQ connected to your Azure AI Search indices.
Foundry IQ's agentic retrieval engine will pull exactly the right context, allowing GPT-5.5 to synthesize the answer and cite the original documentation.
Long Context Handling:
GPT-5.5 models support massive context windows, allowing them to ingest extensive previous conversational turns alongside detailed internal documentation in a single pass without losing track of important rules.
Note:
GPT-5.5 is a large language model (LLM).While it is a flagship LLM built on OpenAI's advanced transformer architecture, it also features natively omnimodal capabilities that allow it to process both text and images seamlessly within a single unified framework. However, when choosing between the specific categories provided, its primary core classification is a Large Language Model (LLM).
Key Details About GPT-5.5
Core Architecture: Large Language Model (LLM) built by OpenAI.
Primary Focus: Highly optimized for complex reasoning, multi-step problem solving, coding, and autonomous agentic workflows.
Input/Output Capabilities: Supports text and image inputs with text-based outputs.


Reference:

https://developers.openai.com/api/docs/guides/reasoning



DRAG DROP (Drag and Drop is not supported)
You have a Microsoft Foundry project that contains a deployed ticket-triage agent.
You discover that sometimes the agent responds without calling any tools, even when a tool is required.
You need to ensure that the agent calls a tool during execution.
How should you complete the Python code? To answer, drag the appropriate values to the correct targets.
Each value may be used once, more than once, or not at all. You may need to drag the split bar between panes or scroll to view content.
Note: Each correct selection is worth one point.
Select and Place:

  1. See Explanation section for answer.

Answer(s): A

Explanation:





Box 1: "tool_choice"
To fix this issue and ensure that the agent always calls a tool during execution, you need to configure the tool_choice parameter in the run_payload.
Box 2: "required"
In Microsoft Foundry's Agent Service (similar to the underlying Azure OpenAI / Assistants API architecture),
setting "tool_choice": "required" forces the model to select and execute one or more tools before returning its response.


Reference:

https://learn.microsoft.com/en-us/azure/foundry/agents/concepts/tool-best-practice



You are building a web app named App1 that generates responses by using a model deployed to a Microsoft
Foundry project named Project1.
Before sending the prompts to the model, App1 must retrieve documents by using Azure AI Search.
You need to integrate Project1 and App1. The solution must meet the following requirements:
▪ Multiple client applications must use the same search configuration.
▪ A security policy must prevent key-based authentication.
▪ Administrative effort must be minimized.
What should you do?

  1. Create a custom HTTP connection in Foundry and manually configure Azure AI Search endpoints per application.
  2. Configure an Azure AI Search connection in Project1 and reference the connection in each application.
  3. Call Azure AI Search directly from each application by using Microsoft Entra authentication.
  4. Enable a managed identity for each application and call Azure AI Search directly.

Answer(s): B

Explanation:

To meet your security and architecture requirements, you must add the Azure AI Search instance as a
Connection within your Azure AI Foundry project and configure Managed Identities for role-based access control (RBAC).
To securely unify your search configuration without API keys, add the Azure AI Search instance as a shared
Connection in your Azure AI Foundry project, disable key authentication on the search service, and authorize your applications using Azure RBAC and Managed Identities.
Note:
*-> 1. Create a Project Connection
Connect Azure AI Search directly inside the Azure AI Foundry hub or project.
*-> Share the same search service configuration across all connected client applications automatically.
Centralize your search endpoint details to reduce administrative overhead.
2. Disable Key Authentication
3. Enable Managed Identities
4. Update the Web App Code


Reference:

https://learn.microsoft.com/en-us/azure/foundry-classic/tutorials/copilot-sdk-create-resources



DRAG DROP (Drag and Drop is not supported)
You have a Microsoft Foundry project that contains an agent used by the financial analysts at your company.
You need to optimize the agent workflow by providing additional data access and processing capabilities. The solution must meet the following requirements:
-▪ Ensure that the agent can perform calculations during conversations.
▪ Ensure that the agent can access up-to-date information from public websites.
▪ Ensure that the agent can retrieve information from documents uploaded directly to the agent.
What should you use for each requirement? To answer, drag the appropriate tools to the correct requirements.
Each tool may be used once, more than once, or not at all. You may need to drag the split bar between panes or scroll to view content.
Note: Each correct selection is worth one point.
Select and Place:

  1. See Explanation section for answer.

Answer(s): A

Explanation:






Box 1: Grounding with Bing Search
Ensure that the agent can access up-to-date information from public websites.
You should use Grounding with Bing Search to optimize your Microsoft Foundry agent workflow.
Large Language Models (LLMs) suffer from fixed knowledge cutoffs and cannot natively access real-time information. Integrating Grounding with Bing Search enables your agent to autonomously fetch up-to-date,
public web data, synthesize facts, and respond with exact inline URL citations.
Box 2: Code interpreter
Ensure that the agent can perform calculations during conversations.
To fulfill this requirement in a Microsoft Foundry Agent Service project, you should use Agent Tools (specifically
Function Calling or a Code Interpreter tool).
In the Microsoft Agent Framework and Foundry Agent Service, tools are the native mechanism used to extend an agent's reasoning capabilities beyond text generation, allowing them to access real-time data and perform exact mathematical computations during a conversation.
Box 3: File Search
Ensure that the agent can retrieve information from documents uploaded directly to the agent.
To give your Microsoft AI Foundry agent the capability to directly reference and retrieve information from uploaded documents, you should enable and configure the File Search tool (via the Foundry Agent Service).
This tool functions alongside the agent's system prompts to perform Retrieval-Augmented Generation (RAG)
directly on your data.


Reference:

https://learn.microsoft.com/en-us/azure/foundry-classic/agents/how-to/tools-classic/bing-code-samples https://learn.microsoft.com/en-us/azure/foundry/agents/concepts/runtime-components https://learn.microsoft.com/en-us/azure/foundry/agents/concepts/tool-catalog



You have a Microsoft Foundry project that contains a prompt agent used by a customer support web app.
The agent is invoked from a Python service that does NOT run in the Foundry portal.
You need to implement end-to-end tracing to capture latency breakdowns and exceptions across agent runs.
Which two components can you use? Each correct answer presents a complete solution.
Note: Each correct selection is worth one point.

  1. a Log Analytics workspace
  2. Application Insights
  3. OpenTelemetry
  4. the Azure Monitor Agent
  5. Microsoft Sentinel

Answer(s): B,C

Explanation:

To implement end-to-end tracing for a Microsoft Foundry prompt agent invoked from an external Python service, you should use a combination of Azure Application Insights, the Azure AI Projects SDK, and
OpenTelemetry.
Core Components
Azure Application Insights: This serves as the centralized telemetry store. You must connect an Application
Insights resource to your Foundry project via the Tracing tab in the Foundry portal to enable ingestion.
Azure AI Projects SDK (azure-ai-projects): Used in your Python service to create an AIProjectClient, which authenticates and communicates with the Foundry agent.
OpenTelemetry Python SDK: Foundry tracing is built on the OpenTelemetry standard. You will need opentelemetry-sdk and azure-monitor-opentelemetry to instrument your external service.


Reference:

https://learn.microsoft.com/en-us/azure/foundry/observability/how-to/trace-agent-setup



You have a customer support agent that uses the Microsoft Foundry Agent Service.
Sometimes, customers return to a session days later to continue the same support case, and the agent must resume with the full historical context. The agent must provide the following:
▪ Multi-turn continuity within the session
▪ Cross-session continuity for the same case
▪ Access to the full interaction history, including user messages, agent messages, tool calls, and tool outputs
You need to ensure that the agent automatically reloads the complete history on each new turn.
What should you do?

  1. Create and reuse a conversation by storing the conversation’s ID and supplying the ID on subsequent requests.
  2. Persist only the final model response stored in the client application and prepend the response to future prompts.
  3. Enable memory summarization on the agent definition to persist the context automatically.

Answer(s): A

Explanation:

To achieve full, multi-turn, and cross-session continuity with complete historical context (including user messages, agent messages, tool calls, and tool outputs), you must leverage the durable Conversation lifecycle provided by the Microsoft Foundry Agent Service runtime.
Because the underlying Agent Service can be stateless at the inference tier, managing continuity across days requires linking your agent interactions to a persistent, server-side conversation ID.
Here is exactly what must be done to meet your requirements:
*-> 1. Maintain a Durable Conversation ID
Instead of creating a brand-new conversation session every time a user returns, your application must persist and map the unique support case to a Microsoft Foundry Conversation ID (typically prefixed with resp_* or conv_*) in your external database.
New Case: Create a new session and capture the generated ConversationId.
Returning Case: Retrieve the previously saved ConversationId associated with that customer's support case from your database.
2. Resume Sessions via the SDK
When the customer returns days later, initialize the agent's session by passing the existing Conversation ID
into the session creation method. This forces the Microsoft Foundry Agent Service to automatically reload the complete, un-summarized thread history on the backend before processing the next turn.


Reference:

https://learn.microsoft.com/en-us/agent-framework/agents/conversations/session



HOTSPOT (Drag and Drop is not supported)
You have a Microsoft Foundry project that contains a deployed chat model.
You have a Python service that sends API requests to the model. The service is integrated with an automated validation system that compares generated outputs against approved response patterns.
Stakeholders report that small wording differences are causing validation mismatches.
You need to update the request parameters to improve output stability. The solution must maximize reasoning quality.
How should you complete the Python code? To answer, select the appropriate options in the answer area.
Note: Each correct selection is worth one point.
Hot Area:

  1. See Explanation section for answer.

Answer(s): A

Explanation:



Box 1: 0
The best configuration for the temperature is 0.
Why Temperature 0 is Best
Maximizes Predictability: Setting temperature to 0 makes the model deterministic.Reduces Mismatches: It always selects the most likely tokens. This minimizes the small wording differences causing your validation failures.
Preserves Reasoning: Lowering temperature does not degrade the core reasoning capability of a model; it simply eliminates random variations in how that reasoning is phrased.
Why Other Options Fail
Temperature 1: This is the default creative balance. It allows standard token variance, which will continue to trigger your pattern-matching mismatches.
Temperature 2: This creates maximum randomness. The output will become chaotic, nonsensical, and completely break both your validation system and reasoning quality.
Box 2: "high"
High output effort is the best configuration for this scenario.
Why High Effort is Best
Maximizes Reasoning: High effort allocates more computational tokens for internal chain-of-thought processing.
Improves Consistency: Better reasoning helps the model strictly adhere to the complex rules and approved response patterns you have defined.
Reduces Edge Cases: Subtle wording deviations usually happen when a model rushes its generation without sufficient planning.


Reference:

https://www.microsoftpressstore.com/articles/article.aspx?p=3192408
https://medium.com/data-science/12-rag-pain-points-and-proposed-solutions-43709939a28c



PDF
Viewing page 4 of 25

Share your comments for Microsoft AI-103 exam with other users:

A
AI Tutor Explanation
6/13/2026 3:10:11 AM

Question 23:
Question 23 describes a multimodal model where users can upload unsafe images that could contain hidden instructions. The goal is to implement controls to mitigate this risk.
Key points to understand

  • Prompt shield for documents: Highly effective. It scans text extracted from inputs (including image text) before it’s sent to the LLM to catch hidden instructions or jailbreaking tries embedded in documents or image-derived text.
  • Prompt shield for user prompts: Partially effective. It blocks direct jailbreak attempts written in the user’s prompt, but doesn’t catch everything, especially content coming from image text.
  • Image moderation: Highly effective. Blocks unsafe or harmful images before they reach the model, preventing many attacks at the source.
  • Protected Material Detection: Not helpful here. It’s designed to detect copyrighted material in outputs, not to protect against inputs that try to manipulate the model.

Why this matters
  • The strongest defense is defense in depth: combine image moderation with both types of prompt shields. The document/text shield catches hidden instructions in extracted image text; the user-prompt shield mitigates jailbreak attempts in user-provided prompts; image moderation stops unsafe images before processing.

On the provided solution note
  • The stated answer (A: “configure a prompt shield for user prompts”) would help, but it alone does not fully meet the goal. A more robust approach is to apply all three controls (document prompt shield, user prompt shield, and image moderation) to achieve stronger risk mitigation.

Bengaluru, India
AI Tutor 👋 I’m here to help!