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.

• ✗

  Analysing customer complaints to identify the root cause of service dissatisfaction

  Why it's wrong here

  Performing root-cause analysis on customer complaints is a diagnostic, retrospective analytics process that seeks to explain why dissatisfaction occurred, often using techniques like Pareto charts or NLP topic clustering. This is fundamentally different from churn prediction, which is a forward-looking supervised learning task that trains on labeled data to classify customers as likely churners or loyalists. Root-cause analysis explains past service failures; it does not generate a risk score for a specific customer's future inactivity.

• ✓

  Using supervised classification to predict which customers are likely to cancel or become inactive

  Why this is correct

  Churn prediction is a canonical supervised classification problem: historical data is labeled with a binary target (churned vs. retained), and a model learns patterns from features like usage frequency, purchase recency, support interactions, and engagement metrics. After training, the model outputs a probability score for each current customer, enabling the business to proactively target the highest-risk individuals with retention offers or interventions. This is exactly the forward-looking, machine-learning-driven approach that defines churn prediction in Azure AI workloads.

• ✗

  Detecting when a customer has already churned based on their last login date

  Why it's wrong here

  Detecting when a customer has already churned is a retrospective, rule-based determination that simply checks whether the last login date exceeds a predefined threshold (e.g., 90 days). This is a deterministic lookup on historical activity, not a predictive model, and it provides no lead time to intervene because the customer has already left. Churn prediction, by contrast, applies supervised classification to historical patterns of engaged versus lost customers to score the likelihood of future churn while there is still time to act.

• ✗

  Using NLP to understand why customers write negative reviews before leaving

  Why it's wrong here

  Using NLP to analyze negative review sentiment is a text-mining task that extracts opinions, emotions, or topics from unstructured text, not a churn-prediction model. Churn prediction relies on structured behavioral and transactional features such as login frequency, purchase history, support tickets, or payment delays, and it forecasts a binary outcome (churn vs. retain). While negative reviews might correlate with churn, NLP alone cannot assign a probability of future cancellation or target at-risk customers in the way a supervised classification model does.

Go deeper

Related to this question