Q01 - Question
A team has listed a long set of potential harms for a customer support copilot that uses a generative model with company documents as customization data. The team has limited time before the next milestone. What should the team do next in the map stage?
Domain: Plan and manage an Azure AI solution (25–30%) Type: Single choice
- A. Prioritize the identified harms by likelihood and potential impact, then test and verify them and document the verified harms.
- B. Apply guardrails to the deployed model so that all listed harms are blocked before any testing.
- C. Replace the customization data with public internet data so that the harm list no longer applies.
- D. Move directly to a phased delivery plan and collect harm reports from early users.
A is correct.
Explanation: Mapping harms includes identifying potential harms, prioritizing them by likelihood and impact, testing and verifying that they can occur, and documenting and sharing the verified harms with stakeholders. Prioritization keeps limited testing effort focused on the harms that matter most.
B is incorrect: Guardrails are a mitigation technique applied after harms are identified and measured, so applying them first leaves the team without verified harms to target or a baseline to compare against.
C is incorrect: Potential harms depend on the services, models, and customization data used, so changing the data source does not remove the need to map harms.
D is incorrect: Phased delivery belongs to the manage stage after mapping, measuring, and mitigating, so releasing first exposes users to harms that were never verified.
Q02 - Question
A retail company wants an application that transcribes recorded customer calls, summarizes them, and extracts product names from scanned receipts. Before any resources are provisioned, what should the solution architect do first?
Domain: Plan and manage an Azure AI solution (25–30%) Type: Single choice
- A. Provision one of every available Azure AI resource so that all options remain open during development.
- B. Write the custom application code first and decide which AI capabilities to add after testing.
- C. Determine the specific AI capabilities the application needs, such as computer speech, natural language processing, and information extraction.
- D. Select the developer tools and SDKs for the project team before defining the application workload.
C is correct.
Explanation: Determining the specific AI capabilities to include, such as generative AI and agents, natural language processing, computer speech, computer vision, and information extraction, helps identify the most appropriate AI services to provision, configure, and use.
A is incorrect: Provisioning every resource does not identify which capabilities the workload requires, and it adds resources you do not need to configure or manage.
B is incorrect: Custom code is only one part of an AI solution, which also combines machine learning models, AI services, and prompt engineering, so the required capabilities should shape the code.
D is incorrect: Tool and SDK selection follows from the capabilities and services you plan to use, not the other way around.
Q03 - Question
A development team deploys a base language model for an internal summarization assistant. Responses are inconsistent in quality and relevance. The team has no labeled training data, no budget for additional infrastructure, and needs to improve results in the current sprint. Which optimization approach fits these constraints?
Domain: Plan and manage an Azure AI solution (25–30%) Type: Single choice
- A. Collect a task-specific dataset and fine-tune the deployed model.
- B. Refine the prompts, including the system message, to improve quality, accuracy, and relevance.
- C. Build a retrieval pipeline over an indexed data source before changing the prompts.
- D. Deploy a second model and route requests between the two deployments.
B is correct.
Explanation: Prompt engineering is the process of designing and refining prompts, including the system message, to improve response quality, accuracy, and relevance. It requires no extra infrastructure and no training data, so it is the approach that matches the team's constraints and timeline.
A is incorrect: Fine-tuning requires a smaller task-specific dataset, which the team does not have.
C is incorrect: Retrieval Augmented Generation adds a retrieval step over a data source, which requires additional setup, and prompt engineering is the starting point for optimization.
D is incorrect: Adding a second deployment does not address prompt quality and adds infrastructure the team cannot fund.
Q04 - Question
An insurance company must extract fields from scanned claim forms and translate the extracted text. The team has no data science staff and wants to avoid training models. What approach meets the requirement?
Domain: Plan and manage an Azure AI solution (25–30%) Type: Single choice
- A. Train a custom machine learning model for document processing and a second model for translation.
- B. Build prompt engineering solutions in custom code to replace document processing and translation.
- C. Create a separate Foundry project for each document type to avoid integration work.
- D. Use Foundry Tools, which provide prebuilt APIs and models for tasks such as document intelligence and translation.
D is correct.
Explanation: Foundry Tools are out-of-the-box prebuilt APIs and models that you can integrate into applications for tasks including text analysis, speech, translation, document intelligence, and content understanding. Applications connect to tool-specific endpoints using project authentication or token-based authentication.
A is incorrect: Training custom models requires data science effort that the team does not have, and prebuilt APIs already cover document intelligence and translation.
B is incorrect: Prompt engineering is a separate part of an AI solution and does not replace the prebuilt document intelligence and translation APIs.
C is incorrect: A project organizes solution assets, but creating multiple projects does not by itself provide document extraction or translation capabilities.
Q05 - Question
A development team needs a single place to organize the models, agents, tools, and knowledge connections for a new AI solution, and the team wants both portal-based and code-based management. What should the team use?
Domain: Plan and manage an Azure AI solution (25–30%) Type: Single choice
- A. A Microsoft Foundry project, managed through the Foundry portal or the SDK.
- B. A separate resource for each model and agent, tracked in a spreadsheet by the team.
- C. A source control repository as the only record of the solution's models and agents.
- D. Prompt engineering documentation stored alongside the application code.
A is correct.
Explanation: Microsoft Foundry projects organize the resources, data, code, and assets of an AI solution, and developers use projects to manage models, agents, tools, and knowledge. You can work with a project through the Foundry portal or through the SDK.
B is incorrect: Manual tracking outside the platform does not provide the project-level access to models, agents, tools, and knowledge connections.
C is incorrect: A repository stores code and files, but it does not provide the project structure that organizes Foundry assets.
D is incorrect: Prompt engineering is one part of an AI solution, and documentation alone does not manage the solution's models, agents, and knowledge connections.
Q06 - Question
A team has already refined the system message and added few-shot examples, but the model still returns responses that vary in tone and output format. The team has a curated set of approved sample responses. Which action should the team take next?
Domain: Plan and manage an Azure AI solution (25–30%) Type: Single choice
- A. Fine-tune the pretrained model on the task-specific dataset of approved responses.
- B. Index the approved responses and retrieve them at query time.
- C. Increase the number of few-shot examples until the format stabilizes.
- D. Replace the system message with a longer list of formatting rules.
A is correct.
Explanation: Fine-tuning takes a pretrained language model and further trains it on a smaller, task-specific dataset to produce behavior consistent with the examples. It applies when prompts and few-shot examples do not reliably achieve the needed style, tone, or format, which is the situation described.
B is incorrect: Retrieval adds factual grounding to the prompt and does not enforce consistent style and format.
C is incorrect: The team already tested few-shot examples and the behavior remains inconsistent.
D is incorrect: The system message was already refined, so further prompt changes are unlikely to deliver the required consistency.
Q07 - Question
A development team deployed a model and must report both subjective response quality and repeatable measures before a release review. Which two Microsoft Foundry evaluation approaches should the team use? (Choose two.)
Domain: Plan and manage an Azure AI solution (25–30%) Type: Multiple choice
- A. Apply the catalog industry filter to confirm the model suits the scenario.
- B. Use manual review of model responses in the playground.
- C. Compare provider names listed on competing model cards.
- D. Run dataset-based evaluations that produce automated generation quality, safety, and NLP metrics.
B and D are correct.
Explanation: Evaluation in Microsoft Foundry combines manual testing with automated metrics. Manual review in the playground captures subjective judgment about responses, and dataset-based evaluations produce automated generation quality, safety, and NLP metrics that you can repeat across runs. The results guide prompt engineering, model choice, RAG integration, or fine-tuning.
A is incorrect: The industry filter helps you discover candidate models in the catalog, but it does not measure how a deployed model performs on your data.
C is incorrect: Provider names on model cards describe model origin and do not measure response quality or safety.
Q08 - Question
An engineer plans to complete a hands-on activity that fine-tunes a language model in Microsoft Foundry. What must the engineer have in place before starting?
Domain: Plan and manage an Azure AI solution (25–30%) Type: Single choice
- A. An existing fine-tuned model deployed in another workspace.
- B. A published retrieval index over customer documents.
- C. A set of validated system messages approved by a security team.
- D. An Azure subscription with administrative access.
D is correct.
Explanation: To fine-tune a language model in Microsoft Foundry as part of the hands-on activity, you need an Azure subscription in which you have administrative access.
A is incorrect: The activity creates the fine-tuned model, so an existing fine-tuned deployment is not a prerequisite.
B is incorrect: A retrieval index supports Retrieval Augmented Generation and is not required to fine-tune a model.
C is incorrect: System message review is part of prompt engineering work and is not stated as a requirement for the fine-tuning activity.
Q09 - Question
A team has documented a set of verified harms for a generative AI solution and now needs to quantify how often the solution produces them. Which two actions should the team take during the measure stage? Each correct answer presents part of the solution.
Domain: Plan and manage an Azure AI solution (25–30%) Type: Multiple choice
- A. Restrict testing to prompts that users already submitted in production.
- B. Submit a diverse set of prompts that are likely to produce each documented harm, collect the model output, and evaluate it against predefined criteria.
- C. Enable a phased delivery plan and use early user reports as the only source of measurement.
- D. Create an initial baseline that quantifies the harms, then repeat automated testing with periodic manual testing and verification.
B and D are correct.
Explanation: Measuring harms means submitting diverse prompts that are likely to produce each documented harm, collecting the model output, and evaluating it against predefined criteria. The results form an initial baseline that you compare against as you apply mitigations, and you scale the work by automating testing while still performing periodic manual testing and verification.
A is incorrect: Limiting tests to prompts that were already submitted does not cover the full range of documented harms and gives an incomplete baseline.
C is incorrect: Phased delivery and user reports belong to the manage stage and cannot replace controlled testing against predefined evaluation criteria before release.
Q10 - Question
A team must shortlist models in the Microsoft Foundry portal for a summarization feature in a regulated industry. They want to narrow the catalog to a manageable set before any testing begins. Which approach in the Foundry Models catalog best supports this first step?
Domain: Plan and manage an Azure AI solution (25–30%) Type: Single choice
- A. Filter the catalog by collection, inference task, fine-tuning method, and industry, then review the model cards of the remaining models.
- B. Deploy every model that mentions summarization and compare the responses in the playground.
- C. Sort the catalog alphabetically by provider and select the first model from each provider.
- D. Request a custom model build because the catalog does not expose responsible AI information.
A is correct.
Explanation: The Foundry Models catalog supports searching and filtering by collection, capabilities, source, inference tasks, fine-tuning methods, and industry. After filtering, each model card shows the provider, capabilities, benchmark metrics, responsible AI considerations, and deployment options, so you can shortlist candidates before testing.
B is incorrect: Deploying every candidate model adds deployment work before you have used the filters and model cards to reduce the candidate list.
C is incorrect: Provider name and alphabetical order are not selection criteria; the catalog filters and model card details are the documented way to narrow the list.
D is incorrect: Model cards already include responsible AI considerations, so you do not need a custom build to see that information.