Microsoft - AI-901: Microsoft Azure AI Fundamentals (beta)
Sample Questions
Question: 55
Measured Skill: Implement AI solutions by using Microsoft Foundry (55–60%)
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
(NOTE: Each correct selection is worth one point.)
| A | System prompts can be used to authorize users: Yes
A system prompt is used to reduce tokens per minute (TPM): Yes
A system prompt guides the behavior of a generative AI model: Yes |
| B | System prompts can be used to authorize users: Yes
A system prompt is used to reduce tokens per minute (TPM): Yes
A system prompt guides the behavior of a generative AI model: No |
| C | System prompts can be used to authorize users: Yes
A system prompt is used to reduce tokens per minute (TPM): No
A system prompt guides the behavior of a generative AI model: Yes |
| D | System prompts can be used to authorize users: No
A system prompt is used to reduce tokens per minute (TPM): Yes
A system prompt guides the behavior of a generative AI model: No |
| E | System prompts can be used to authorize users: No
A system prompt is used to reduce tokens per minute (TPM): No
A system prompt guides the behavior of a generative AI model: Yes |
| F | System prompts can be used to authorize users: No
A system prompt is used to reduce tokens per minute (TPM): No
A system prompt guides the behavior of a generative AI model: No |
Correct answer: EExplanation:
System messages (sometimes called a system prompt or metaprompt) help you steer an Azure OpenAI chat model toward the behavior, tone, and output format you want.
A system message is a set of instructions and context you provide to the model to guide its responses. You typically use it to:
- Define the assistant’s role and boundaries.
- Set tone and communication style.
- Specify output formats (for example, JSON).
- Add safety and quality constraints for your scenario.
A system message can be one short sentence:
You are a helpful AI assistant.
Or it can be multiple lines with structured rules and formatting requirements.
Question: 56
Measured Skill: Implement AI solutions by using Microsoft Foundry (55–60%)
Select the answer that correctly completes the sentence.
| A | A keyword list defines which fields to extract when analyzing content. |
| B | Optical character recognition (OCR)-only processing defines which fields to extract when analyzing content. |
| C | A schema defines which fields to extract when analyzing content. |
| D | A synchronous API call defines which fields to extract when analyzing content. |
Correct answer: CExplanation:
An analyzer in Azure Content Understanding in Foundry Tools is a configurable processing unit that defines how your content is analyzed and what information is extracted. An analyzer uses field schemas to guide the extraction process, specifying exactly what information, e.g., names, dates, amounts the system should identify.
Reference: What is a Content Understanding analyzer?
Question: 57
Measured Skill: Implement AI solutions by using Microsoft Foundry (55–60%)
For each of the following statements, select Yes if the statement is true. Otherwise, select No.
(NOTE: Each correct selection is worth one point.)
| A | In the new Microsoft Foundry portal, you must fine-tune a model before you can deploy the model: Yes
In the new Microsoft Foundry portal, you can test a model from the model catalog only after you deploy the model: Yes
In the new Microsoft Foundry portal, you can deploy a model from the model catalog only after retraining the model: Yes |
| B | In the new Microsoft Foundry portal, you must fine-tune a model before you can deploy the model: Yes
In the new Microsoft Foundry portal, you can test a model from the model catalog only after you deploy the model: Yes
In the new Microsoft Foundry portal, you can deploy a model from the model catalog only after retraining the model: No |
| C | In the new Microsoft Foundry portal, you must fine-tune a model before you can deploy the model: Yes
In the new Microsoft Foundry portal, you can test a model from the model catalog only after you deploy the model: No
In the new Microsoft Foundry portal, you can deploy a model from the model catalog only after retraining the model: Yes |
| D | In the new Microsoft Foundry portal, you must fine-tune a model before you can deploy the model: No
In the new Microsoft Foundry portal, you can test a model from the model catalog only after you deploy the model: Yes
In the new Microsoft Foundry portal, you can deploy a model from the model catalog only after retraining the model: No |
| E | In the new Microsoft Foundry portal, you must fine-tune a model before you can deploy the model: No
In the new Microsoft Foundry portal, you can test a model from the model catalog only after you deploy the model: Yes
In the new Microsoft Foundry portal, you can deploy a model from the model catalog only after retraining the model: Yes |
| F | In the new Microsoft Foundry portal, you must fine-tune a model before you can deploy the model: No
In the new Microsoft Foundry portal, you can test a model from the model catalog only after you deploy the model: No
In the new Microsoft Foundry portal, you can deploy a model from the model catalog only after retraining the model: No |
Correct answer: FExplanation:
Microsoft Foundry is a unified Azure platform-as-a-service offering for enterprise AI operations, model builders, and application development. This foundation combines production-grade infrastructure with friendly interfaces, enabling developers to focus on building applications rather than managing infrastructure.
Microsoft Foundry unifies agents, models, and tools under a single management grouping with built-in enterprise-readiness capabilities including tracing, monitoring, evaluations, and customizable enterprise setup configurations. The platform provides streamlined management through unified role-based access control (RBAC), networking, and policies under one Azure resource provider namespace.
You can deploy base (prebuilt) models directly from the model catalog without any fine-tuning. Fine-tuning is optional, used only for customization.
In Azure AI Foundry, you can test models in the playground before deployment. Deployment is required only to use the model in applications.
Retraining/fine-tuning is not a prerequisite for deployment.
References:
What is Microsoft Foundry?
Choose and deploy models from the model catalog in Microsoft Foundry portal
Microsoft Foundry Playgrounds
Question: 58
Measured Skill: Implement AI solutions by using Microsoft Foundry (55–60%)
You are developing an application that analyzes invoices by using Azure Content Understanding in Foundry Tools.
You need to ensure that the application retrieves the analysis results after processing completes.
How should you complete the Python code?
(To answer, select the appropriate option in the answer area. NOTE: Each correct selection is worth one point.)
| A | P1: get_result() |
| B | P1: result() |
| C | P1: status |
| D | P1: wait |
Correct answer: BExplanation:
In Azure AI SDKs, including Document Intelligence, begin_analyze() returns a poller object. The poller represents a long-running operation (LRO).
The poller.result() method returns the result of the long running operation, or the result available after the specified timeout.
Reference: AnalyzeLROPoller Class
Question: 59
Measured Skill: Identify AI concepts and responsibilities (40–45%)
Select the answer that correctly complete the sentence.
| A | The Model catalog is used for comparing and deploying a wide range of models for generative AI development in Microsoft Foundry. |
| B | The Monitor page is used for comparing and deploying a wide range of models for generative AI development in Microsoft Foundry. |
| C | The Service endpoints page is used for comparing and deploying a wide range of models for generative AI development in Microsoft Foundry. |
| D | The Solution templates page is used for comparing and deploying a wide range of models for generative AI development in Microsoft Foundry. |
Correct answer: AExplanation:
The model catalog in Foundry portal is the hub for discovering and using a wide range of models to build generative AI applications. The model catalog features hundreds of models across model providers like Azure OpenAI, Mistral, Meta, Cohere, NVIDIA, and Hugging Face, including models that Microsoft trained. Models from providers other than Microsoft are Non-Microsoft Products as defined in Microsoft Product Terms and are subject to the terms provided with the models.
Search and discover models that meet your needs through keyword search and filters. The model catalog also offers the model performance leaderboard and benchmark metrics for select models. Access them by selecting View leaderboard and Compare models. Benchmark data is also available from the model card's Benchmarks tab.
Some of the filters available in the model catalog are:
- Collection: Filter models based on the model provider collection.
- Industry: Filter for the models that are trained on industry-specific dataset.
- Capabilities: Filter for unique model features like reasoning and tool calling.
- Inference tasks: Filter models based on the inference task type.
Some of the details available in the model card are:
- Quick facts: Key information about the model at a quick glance
- Details tab: Detailed information about the model, like description, version info, and supported data type
- Benchmarks tab: Performance benchmark metrics for select models
- Deployments tab: A list of existing deployments for the model
- License tab: Legal information related to model licensing
References:
Microsoft Foundry Models overview
Choose and deploy models from the model catalog in Microsoft Foundry portal