AI Associate Data for AI Practice Test 2 — 15 Questions

Question 1

Refer to the exhibit. The data pipeline is failing. What is the most likely cause?

Accepted Answer

Schema mismatch between source and target.. A schema mismatch between source and target is the most common cause of pipeline failures in data integration workflows within Salesforce Data Cloud. When the source data structure (e.g., field names, data types, or object relationships) does not match the target data model (e.g., a Data Cloud data lake object or Einstein Studio dataset), the pipeline cannot map or transform the data correctly, leading to errors during ingestion or transformation stages. This is especially relevant in Salesforce Data Cloud's data pipelines where schema validation is enforced at runtime.

Answer

Network timeout.

Answer

Missing required field in source data.

Answer

Insufficient memory.

Question 2

Refer to the exhibit. A developer runs a SOQL query. What does the output indicate?

Accepted Answer

The query returned 10 records in total.. The SOQL query output shows '10 records returned' with no error or partial result indicator, confirming that the query completed successfully and returned exactly 10 records. In Salesforce SOQL, the query result includes a 'totalSize' field that reflects the total number of records matching the query criteria, and here it matches the number of records displayed, indicating a complete and successful retrieval.

Answer

The query is still processing.

Answer

The output is incomplete.

Answer

The query failed.

Question 3

A retail company uses Einstein Next Best Action with customer data from Data Cloud. The recommendations are not personalized. The admin checks the data quality dashboard and finds that the 'Customer_Profile' object has 40% records with missing 'PreferredChannel' field. What is the best course of action?

Accepted Answer

Impute the missing values using the mode of the field.. Imputing missing values using the mode (most frequent value) of the 'PreferredChannel' field is a standard data preprocessing technique that preserves the dataset size and statistical distribution. In Einstein Next Best Action, missing categorical data can degrade model personalization, and mode imputation is a simple, effective way to handle this without losing records or altering the model structure.

Answer

Remove the field from the model.

Answer

Increase the data refresh frequency.

Answer

Train the model with only records that have non-null PreferredChannel.

Question 4

A company wants to use Einstein Reply Recommendations in Service Cloud. What data is required to train the model?

Accepted Answer

Case comments and chat transcripts.. Einstein Reply Recommendations in Service Cloud uses historical service interactions to suggest relevant replies to agents. The model is trained on case comments and chat transcripts because these contain the natural language patterns and resolutions that agents actually use in real-time service conversations, enabling the AI to learn effective response strategies.

Answer

Knowledge articles only.

Answer

Email templates.

Answer

Historical email replies and customer satisfaction ratings.

Question 5

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?

Accepted Answer

Use oversampling techniques.. 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.

Answer

Remove the churned records.

Answer

Increase the amount of training data.

Answer

Change the prediction field.

Question 6

A data scientist is using Einstein Discovery to analyze sales data. The model results show a high correlation between two predictor variables. Which TWO actions should the data scientist take?

Accepted Answer

Combine them into a single feature.. Option B is correct because when two predictor variables are highly correlated (multicollinearity), combining them into a single engineered feature (for example, via averaging, ratio, or PCA-style transformation) eliminates the redundancy while preserving the shared signal, giving the model a cleaner, non-duplicated input. Option E is correct because removing one of the correlated variables is the standard, simplest remedy for multicollinearity: since the two predictors carry largely overlapping information, dropping one reduces variance inflation and stabilizes the model's coefficient estimates without meaningful loss of predictive power. Option A is not appropriate here because regularization (L1/L2) mitigates overfitting and shrinks coefficients but does not resolve the underlying correlation/redundancy between predictors. Option C is wrong because including both correlated variables inflates variance and destabilizes importance/coefficient estimates rather than adding genuine information. Option D is wrong because increasing sample size does not fix multicollinearity — the correlation between the predictors persists regardless of how many rows are added.

Answer

Apply regularization.

Answer

Include both to capture more information.

Answer

Increase the sample size.

Question 7

Refer to the exhibit. What effect does this masking policy have on the data used for training an Einstein model?

Accepted Answer

SSN and CreditCard fields are partially masked, showing only the last four characters.. The masking policy in Einstein applies a partial mask to sensitive fields like SSN and CreditCard, showing only the last four characters while obscuring the rest. This ensures that the data used for training retains its structural utility for model learning without exposing full sensitive values, which is why option D is correct.

Answer

Only SSN is masked.

Answer

SSN and CreditCard fields are encrypted.

Answer

SSN and CreditCard fields are completely removed from training data.

Question 8

Refer to the exhibit. What is the most likely cause of this error?

Accepted Answer

A division formula in the data transform is dividing by a field that contains zero values.. The error shown in the exhibit is a division-by-zero runtime error, which occurs when a formula in a data transform attempts to divide a value by a field that contains zero. In Salesforce Data Cloud, data transforms execute field-level calculations, and if a divisor field holds a zero, the system throws a 'Divide by zero' exception. Option D correctly identifies this as the most likely cause because the error message explicitly indicates a division operation failed due to a zero divisor.

Answer

The source data is missing required fields.

Answer

The data transform has a recursive formula.

Answer

The target field expects an integer but source provides null.

Question 9

Refer to the exhibit. What data quality issue does the exhibit reveal?

Accepted Answer

The Fall Clearance campaign has no response data.. The exhibit reveals that the Fall Clearance campaign has no response data. In data quality analysis, missing response data for a campaign indicates an incomplete dataset, which can skew analysis and model training. This is a data quality issue because it affects the completeness and reliability of the data.

Answer

The Summer Sale campaign has duplicate records.

Answer

The query syntax is incorrect.

Answer

The data is not normalized.

Question 10

A company is training a customer service chatbot using historical conversation logs. Which TWO data preparation practices should be followed to ensure data quality?

Accepted Answer

Include answers with varied phrasing to enhance language variety. Option B is correct because including answers with varied phrasing exposes the chatbot to diverse linguistic patterns, synonyms, and sentence structures, which improves its ability to generalize and respond naturally to different user inputs. Option E is correct because complete transcripts provide full conversational context—both user turns and agent responses—so the model can learn proper dialogue flow, intent resolution, and turn-taking rather than being trained on truncated or ambiguous exchanges. Option A is not among the correct answers; while privacy protection is important, excluding all user identifiers is a compliance/anonymization concern rather than a data-quality practice for improving chatbot training. Option C is incorrect because using only successfully resolved interactions introduces selection bias and omits valuable examples of failed or escalated conversations that help the model learn edge cases. Option D is incorrect because filtering to only English conversations reduces linguistic diversity and limits the chatbot's coverage, which is a scope decision rather than a data-quality improvement.

Answer

Exclude all user identifiers to protect privacy

Answer

Include only successful interactions that were resolved

Answer

Filter only English conversations for consistency

Question 11

Before training an Einstein Prediction model, a data analyst must perform data quality checks. Which THREE checks are most critical?

Accepted Answer

Remove duplicate records that could cause data leakage. Option B is correct because duplicate records can cause data leakage between training and test splits, artificially inflating model accuracy and producing an unreliable Einstein Prediction model. Option C is correct because inconsistent data types (for example, dates stored as strings in some rows and Date objects in others) break feature engineering and cause the model training job to fail or misinterpret fields. Option E is correct because missing values in key fields degrade prediction quality and must be identified and handled (imputed or excluded) before training. Option A is not among the marked correct answers, and option D is incorrect because Einstein Prediction models such as gradient-boosted trees do not require features to follow a normal distribution.

Answer

Confirm that label distribution matches the target baseline

Answer

Ensure all features follow a normal distribution

Question 12

Refer to the exhibit. A data file for click-through model training has the above content. Which data quality issue is most critical to address before training?

Accepted Answer

Missing value in the Conversions column for the third row. Missing values in the Conversions column directly impact the supervised learning target variable. If the label (conversion) is missing for a training instance, the model cannot learn the correct mapping from features to outcome, leading to biased or incomplete training. This is a critical data quality issue that must be addressed before training, typically via imputation or row removal.

Answer

The header row is missing a column name for the last field

Answer

Inconsistent date formats across rows

Answer

Clicks column is an integer but may need scaling

Question 13

Refer to the exhibit. A data analyst runs a profile on a dataset and sees these statistics. Based on best practices, which action should be taken first?

Accepted Answer

Remove the 200 duplicate records. Duplicate records introduce bias and redundancy, leading to overfitting or skewed model performance. Removing duplicates is a standard first step in data preprocessing to ensure data integrity before handling missing values or outliers. In the context of the AI Associate exam, best practices prioritize deduplication early in the data cleaning pipeline.

Answer

Impute the 500 missing values with the mean

Answer

Remove the 50 outliers in the Amount field

Answer

Skip all preprocessing and train the model directly

Question 14

For an AI project, data must be stored in a way that supports both training and real-time inference. Which storage solution meets this requirement?

Accepted Answer

Data lake (e.g., Amazon S3 or Azure Data Lake). A data lake (e.g., Amazon S3 or Azure Data Lake) is the correct choice because it can store vast amounts of raw, unstructured, and structured data in its native format, making it ideal for training AI models on diverse datasets. At the same time, data lakes support real-time inference by enabling direct access to data via APIs or streaming services (e.g., AWS Lambda or Azure Functions) without the latency of transforming data into a schema-on-write structure. This dual capability—handling both batch processing for training and low-latency reads for inference—is a key requirement that other storage solutions cannot fulfill as effectively.

Answer

Data warehouse (e.g., Snowflake)

Answer

Relational database (e.g., PostgreSQL)

Answer

In-memory cache (e.g., Redis)

Question 15

A data engineer needs to create a feature that represents the average purchase amount per customer over the last 30 days. The transactional data is timestamped. Which feature engineering technique is most appropriate?

Accepted Answer

Rolling average of purchase amounts over a 30-day window. A rolling average over a 30-day window directly computes the average purchase amount per customer for only the most recent 30 days of transactions, which matches the requirement of a time-sensitive feature. This technique uses a sliding window function (e.g., AVG() with a ROWS or RANGE frame in SQL, or rolling().mean() in pandas) that respects the timestamp order, ensuring only relevant data contributes to the feature.

Answer

Sum of all purchase amounts per customer

Answer

Count of purchases per customer

Answer

Minimum purchase amount per customer