AI-900 Sample Questions & Answers
Generative AI services and capabilities get the closest attention, just ahead of common AI workloads and the principles behind responsible AI, machine learning techniques and what Azure Machine Learning offers, and computer vision solution types.
Launch the full AI-900 simulator →Showing 10 of 20 free samples.
- Question 1Beginner
Describe features of Natural Language Processing (NLP) workloads on Azure · Identify features of common NLP Workload Scenarios
An e-commerce company is building a chatbot using the Azure Bot Service and Azure AI Language's conversational language understanding (CLU) feature. The goal is to handle customer inquiries about order status. A user might ask, "Where is my shipment with ID 789123?". In this context, what is the 'intent'?
Show answer & explanation
Correct answer: C
In conversational AI, the 'intent' represents the user's underlying goal or intention. In this case, the user's goal is to find out the status of their order. The specific question is the 'utterance', and '789123' is an 'entity'.
- Question 2Beginner
Describe Artificial Intelligence workloads and considerations · Identify guiding principles for responsible AI
A city's public transportation authority is developing an AI model to predict bus arrival times. The model is trained on historical data but shows significant bias, consistently underestimating travel times for routes in low-income neighborhoods. This issue violates which principle of Microsoft's Responsible AI?
Show answer & explanation
Correct answer: B
The principle of Fairness is violated. This principle dictates that AI systems should treat all people fairly and avoid affecting similarly situated groups of people in different ways. The model's systemic underperformance for a specific demographic (people in low-income neighborhoods) is a clear example of algorithmic bias and unfairness.
- Question 3Intermediate
Describe Artificial Intelligence workloads and considerations · Identify guiding principles for responsible AI
To help developers understand why an automated machine learning (AutoML) model in Azure Machine Learning makes certain predictions, the platform provides model explanations. This capability directly supports which Microsoft Responsible AI principle?
mindmap root((Responsible AI)) Fairness Reliability & Safety Privacy & Security Inclusiveness Accountability (Transparency) Explainability InterpretabilityShow answer & explanation
Correct answer: D
Model explanations, which help humans understand the reasoning behind an AI's decisions, are a core component of the Transparency principle. Transparency is about ensuring that AI systems are understandable, and providing tools for interpretability and explainability is key to achieving this.
- Question 4Beginner
Describe Artificial Intelligence workloads and considerations · Identify features of common AI workloads
Which type of AI workload is primarily concerned with identifying data points, events, or observations that deviate from a system's normal behavior?
Show answer & explanation
Correct answer: B
Anomaly detection is the AI workload focused on identifying rare items, events, or observations which raise suspicions by differing significantly from the majority of the data. Examples include fraud detection in credit card transactions or identifying a faulty sensor in an IoT network.
- Question 5Intermediate
Describe fundamental principles of machine learning on Azure · Describe core machine learning concepts
A data scientist is training a classification model to predict customer churn. After training, they evaluate the model on the test dataset and notice that it performs exceptionally well on the data it was trained on, but its accuracy is very poor on the unseen test data. What is the most likely term for this phenomenon?
Show answer & explanation
Correct answer: C
Overfitting occurs when a machine learning model learns the training data too well, including its noise and random fluctuations. This results in a model that performs very well on the training data but fails to generalize to new, unseen data, leading to poor performance on the test set.
- Question 6Beginner
Describe fundamental principles of machine learning on Azure · Identify common machine learning techniques
An agricultural company is building a machine learning model to predict the annual yield of corn (in tons per acre) based on features like rainfall, temperature, soil type, and fertilizer usage. Which type of machine learning model is most appropriate for this task?
Show answer & explanation
Correct answer: C
Regression is the correct choice because the goal is to predict a continuous numerical value (corn yield in tons per acre). Classification predicts a category, and clustering groups data without predefined labels.
- Question 7Intermediate
Describe fundamental principles of machine learning on Azure · Identify common machine learning techniques
A marketing team wants to segment its customer base into distinct groups based on purchasing behavior without any predefined labels for the groups. The goal is to discover natural groupings of customers to target with different campaigns. What machine learning technique should be used?
graph TD subgraph Unsupervised_Learning A[Raw Data] --> B{Clustering Algorithm} B --> C[Group 1] B --> D[Group 2] B --> E[Group 3] endShow answer & explanation
Correct answer: B
Clustering is an unsupervised learning technique used to group similar data points together based on their features. Since the marketing team does not have predefined labels for customer groups and wants to discover these groups, clustering is the appropriate method.
- Question 8Advanced
Describe features of Natural Language Processing (NLP) workloads on Azure · Identify Azure tools and services for NLP workloads
Case Study:
Contoso Pharmaceuticals is launching a global medical information service. They need to build a robust AI-powered system to support healthcare professionals worldwide. The system will ingest medical journals, clinical trial results, and drug information sheets from various sources.
Current Situation:
Contoso has a large, unstructured repository of documents in multiple languages, including English, German, and Japanese. Medical professionals need to be able to ask complex questions in their native language and receive concise, accurate answers that are sourced directly from the ingested documents. The current manual research process is slow and inefficient.Requirements:
- The system must understand questions posed in natural language.
- It needs to search the vast repository of documents to find relevant information.
- The answers provided must be synthesized from the source material, not just a list of links.
- The system must be able to translate queries and answers between the supported languages.
- It must identify key medical terms, such as drug names and conditions, within the documents for better indexing.
Question:
Which combination of Azure AI services would best fulfill all of Contoso's requirements for their medical information service?Show answer & explanation
Correct answer: C
This combination directly addresses all the requirements. Azure AI Language's question answering feature can understand queries and synthesize answers from documents. Its entity recognition feature can identify key medical terms (Req 1, 2, 3, 5). Azure AI Translator provides the necessary multilingual capabilities for both queries and answers (Req 4).
- Question 9Intermediate
Describe features of generative AI workloads on Azure · Identify features of generative AI solutions
A developer is using the Azure OpenAI Service to build an application that generates marketing copy. They are using the Chat Completions API. To ensure the model's responses are creative and varied, which parameter should they adjust?
Show answer & explanation
Correct answer: B
The 'temperature' parameter controls the randomness of the model's output. A higher temperature (e.g., 0.8) makes the output more random and creative, while a lower temperature (e.g., 0.2) makes it more deterministic and focused. For creative marketing copy, increasing the temperature is the appropriate action.
- Question 10Beginner
Describe fundamental principles of machine learning on Azure · Describe core machine learning concepts
When a developer uses Azure Machine Learning designer to create a training pipeline, they split the dataset into two parts. One part is used to train the model, and the other part is used to score the trained model and evaluate its performance. What is the second part of the dataset, used for evaluation, commonly called?
Show answer & explanation
Correct answer: B
In machine learning, the data is typically split into a training dataset, used to fit the model, and a validation or test dataset, which is held back and used to provide an unbiased evaluation of the model's performance on unseen data.
Ready for the real thing?
The full AI-900 simulator has every exam-style question, timed mode, and instant scoring.