AI-900 Practice Question: Describe Artificial Intelligence workloads and considerations

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  An AI that recommends Azure pricing tiers based on an organisation's usage patterns

  Why it's wrong here

  Azure Advisor does use usage telemetry to suggest pricing tiers and reserved-capacity purchases, but that is a cost-optimization service, not a recommendation system in the AI-900 sense. A true recommendation system builds a model of each user's preferences from past behavior and ranks content or products for that individual. Advisor's output is an operational/financial suggestion based on static rules and metrics, not a personalized prediction of what a user would most enjoy or need.

• ✓

  AI that predicts user preferences to suggest relevant products, content, or connections

  Why this is correct

  This is the defining behavior of a recommendation system: it predicts individual user preferences from historical behavior such as clicks, purchases, ratings, or views, and then suggests items likely to be relevant. Collaborative filtering, content-based filtering, and hybrid approaches are common techniques, with classic examples including Netflix's movie suggestions, Spotify's Discover Weekly, and Amazon's 'Customers who bought this also bought.' The key is that the model's output is a personalized, user-specific ranking of content or products, not a generic rule-based operational suggestion.

• ✗

  A system that recommends when to retrain an AI model based on performance metrics

  Why it's wrong here

  Triggering model retraining when performance metrics degrade is an MLOps practice involving drift detection, data quality monitoring, and scheduled pipeline automation. The system is making an operational decision about the model lifecycle, not a user-facing suggestion. Its output is an action (retrain or redeploy) rather than a personalized ranking of products, content, or connections for an end user, so it falls outside the 'recommendation system' workload type in the AI-900 syllabus.

• ✗

  AI that recommends the best cloud architecture for a software application

  Why it's wrong here

  Recommending the best cloud architecture is a solution-design activity usually performed by the Azure Well-Architected Review, which evaluates workload requirements such as scalability, reliability, and security. This is an expert/advisory process driven by architectural principles, not a machine learning system that learns user preferences. While it does produce recommendations, it lacks the core recommender mechanism: predicting an individual user's affinity for unseen items based on historical interaction patterns.

Go deeper

Related to this question