Quick Answer
The answer is to use oversampling techniques. When a dataset has severe class imbalance, like only 2% churned records, the model becomes biased toward predicting the majority class, resulting in low accuracy despite high apparent performance. Oversampling, such as SMOTE or random oversampling, artificially increases the minority class records in the training set, allowing Einstein Prediction Builder to learn patterns for churn more effectively. On the Salesforce AI Associate exam, this scenario tests your understanding of how class imbalance impacts predictive models and the specific remedy within Einstein Prediction Builder. A common trap is to assume adding more data overall or adjusting thresholds will fix the bias, but oversampling directly addresses the skewed distribution. Remember the memory tip: “When minority is tiny, oversampling makes it shiny.”
This AI Associate practice question tests your understanding of data for ai. Read the scenario carefully and evaluate each option against the stated constraints before committing to an answer. After answering, compare your reasoning against the explanation and wrong-answer breakdown below. Once you have made your selection, read the full explanation to reinforce the concept and understand why each distractor is designed to mislead on exam day.
A company uses Einstein Prediction Builder to predict customer churn. The model's accuracy is low. The admin reviews the training data and notices that only 2% of records are churned. What should the admin do to improve the model?
Remove the churned records.
Why wrong: Removing the minority class would make the problem worse.
Increase the amount of training data.
Why wrong: More data may not fix the imbalance ratio.
Use oversampling techniques.
Oversampling balances the classes and improves model sensitivity.
Change the prediction field.
Why wrong: Changing the target does not address imbalance.
Answer choices
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
Use oversampling techniques.
Option C is correct because when a dataset has severe class imbalance (only 2% churned records), the model becomes biased toward predicting the majority class (non-churned), leading to low accuracy despite high apparent performance. Oversampling techniques, such as SMOTE or random oversampling, artificially increase the number of churned records in the training set to balance the classes, allowing Einstein Prediction Builder to learn patterns for the minority class more effectively.
Key principle: Answer the scenario, not the keyword: identify the specific constraint before choosing the most familiar-sounding option.
Answer analysis
For each option: why learners choose it and why it is or isn't the right answer here.
Remove the churned records.
Why it's wrong here
Removing the minority class would make the problem worse.
Increase the amount of training data.
Why it's wrong here
More data may not fix the imbalance ratio.
Use oversampling techniques.
Why this is correct
Oversampling balances the classes and improves model sensitivity.
Related concept
Read the scenario before looking for a memorised answer.
Change the prediction field.
Why it's wrong here
Changing the target does not address imbalance.
Common exam traps
Salesforce often tests the misconception that adding more data always improves model performance, but here the trap is that candidates overlook class imbalance and choose 'Increase the amount of training data' (Option B) without realizing that more data with the same imbalance does not solve the problem.
Detailed technical explanation
Under the hood, Einstein Prediction Builder uses automated machine learning (AutoML) with gradient-boosted trees or logistic regression, both of which optimize for overall accuracy and can be overwhelmed by majority-class examples. Oversampling techniques like SMOTE generate synthetic churn examples by interpolating between existing minority-class instances, which helps the model learn decision boundaries for churn without simply duplicating records. In a real-world scenario, a telecom company with 2% churn might use oversampling to achieve a 50/50 split, then evaluate using precision-recall curves rather than accuracy to avoid misleading metrics.
Key takeaway
Answer the scenario, not the keyword: identify the specific constraint before choosing the most familiar-sounding option.
Real-world example
A small business has 20 workstations on the 192.168.1.0/24 network and one public IP from its ISP. The router uses PAT (NAT overload) so all 20 devices share one public address using different source ports. NAT questions test whether you understand the four address terms and which direction each translation applies.
What to study next
Identify which exam domain this question belongs to, review the core concept, then practise similar questions from the same domain.
Data for AI — study guide chapter
Learn the concepts, then practise the questions
Data for AI practice questions
Targeted practice on this topic area only
All AI Associate questions
506 questions across all exam domains
Salesforce AI Associate AI Associate study guide
Full concept coverage aligned to exam objectives
AI Associate practice test guide
How to use practice tests most effectively before exam day
Related practice questions
Use these pages to review the topic behind this question. This is how one missed question becomes focused revision.
Practise AI Associate questions linked to AI Fundamentals.
Practise AI Associate questions linked to AI Capabilities in CRM.
Practise AI Associate questions linked to Ethical Considerations of AI.
Practise AI Associate questions linked to Data for AI.
Practise AI Associate questions linked to AI Associate fundamentals.
Practise AI Associate questions linked to AI Associate scenario.
Practise AI Associate questions linked to AI Associate troubleshooting.
Practice this exam
Short sessions build daily habit. Longer sessions build exam-day stamina. Try a timed session to simulate real conditions.
FAQ
Data for AI — This question tests Data for AI — Read the scenario before looking for a memorised answer..
The correct answer is: Use oversampling techniques. — Option C is correct because when a dataset has severe class imbalance (only 2% churned records), the model becomes biased toward predicting the majority class (non-churned), leading to low accuracy despite high apparent performance. Oversampling techniques, such as SMOTE or random oversampling, artificially increase the number of churned records in the training set to balance the classes, allowing Einstein Prediction Builder to learn patterns for the minority class more effectively.
Identify which exam domain this question belongs to, review the core concept, then practise similar questions from the same domain.
Read the scenario before looking for a memorised answer.
About these practice questions
Courseiva creates original exam-style practice questions with explanations and wrong-answer analysis. It does not publish real exam questions, exam dumps, or protected exam content. Learn why practice questions differ from exam dumps →
Same concept, more angles
These questions test the same concept from different angles. Work through them to make sure you can recognise it however the exam phrases it.
Variation 1. A fraud detection model is being trained on transaction data where only 1% of transactions are fraudulent. The current model predicts 'non-fraud' for all transactions, achieving 99% accuracy. Which technique should be applied to improve model performance?
easyWhy D: Oversampling or undersampling addresses class imbalance, allowing the model to learn minority patterns. Using more features alone doesn't fix imbalance, setting a lower threshold may help but is less common than resampling, and removing minority class is counterproductive.
Last reviewed: Jun 30, 2026
This AI Associate practice question is part of Courseiva's free Salesforce certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI Associate exam.
Question Discussion
Share a tip, memory trick, or ask about the reasoning behind this question. Do not post real exam questions, leaked content, braindumps, or copyrighted exam material. Comments are moderated and may be removed without notice.
Sign in to join the discussion.