AI Associate Practice Test 3 — 25 Questions

Question 1

Which TWO practices are recommended when using AI for automated decision-making in hiring?

Accepted Answer

Regularly audit the model for adverse impact.. Regular auditing for adverse impact is a core ethical practice to detect and mitigate bias in AI-driven hiring systems. Audits involve statistical analysis (e.g., the four-fifths rule from the Uniform Guidelines on Employee Selection Procedures) to compare selection rates across protected groups, ensuring the model does not disproportionately disadvantage certain demographics.

Answer

Use the AI model as the sole decision-maker.

Answer

Use all available data including protected attributes.

Answer

Ignore adverse impact if the model is accurate.

Question 2

An AI Associate deploys an Einstein Bot that uses sentiment analysis to escalate frustrated customers. After launch, the bot escalates disproportionately for non-native English speakers. What is the most likely cause?

Accepted Answer

The sentiment model was trained on a non-representative dataset.. The sentiment analysis model likely exhibits bias due to training data that does not adequately represent the linguistic patterns, idioms, or expressions of non-native English speakers. This causes the model to misinterpret neutral or positive statements from these users as negative or frustrated, leading to disproportionate escalations. A non-representative dataset is a common source of algorithmic bias in AI systems.

Answer

The bot is routing to the wrong department.

Answer

The escalation threshold is set too low.

Answer

The bot is not properly connected to the escalation queue.

Question 3

Which THREE factors should an AI Associate consider when evaluating a model for potential bias?

Accepted Answer

Whether features are correlated with protected attributes.. If a feature is correlated with a protected attribute (e.g., race, gender, age), the model may inadvertently learn and perpetuate discriminatory patterns, even if the protected attribute itself is not used as an input. This is a key source of indirect or proxy bias in machine learning systems.

Answer

The complexity of the model architecture.

Answer

The date the model was last deployed.

Question 4

An administrator is configuring a Salesforce AI model that uses historical sales data. The data includes fields like 'Amount', 'Close_Date', and 'Lead_Source'. What is the primary purpose of data preprocessing in this context?

Accepted Answer

To clean and transform data into a format suitable for model training. Data preprocessing is essential for AI models because raw historical sales data often contains missing values, inconsistent formats, and noise. Cleaning (e.g., handling nulls in 'Amount') and transforming (e.g., encoding 'Lead_Source' into numerical features) ensure the model can learn patterns effectively, directly impacting training accuracy and convergence.

Answer

To generate visualizations for business stakeholders

Answer

To increase the storage capacity of the database

Answer

To enforce data access permissions for different user roles

Question 5

A company wants an AI chatbot that can handle customer inquiries about order status. Which tool should be configured?

Accepted Answer

Einstein Bot. Einstein Bot is the correct tool because it is Salesforce's native AI-powered chatbot designed to handle customer inquiries, including order status, through natural language processing and automated conversations. It can be configured to answer common questions, escalate complex issues, and integrate with backend systems to retrieve real-time order data without human intervention.

Answer

Queues

Answer

Omni-Channel

Answer

Case Assignment Rules

Question 6

Which data transformation is most appropriate for converting categorical variables into numerical format for a machine learning model?

Accepted Answer

One-hot encoding.. One-hot encoding is the correct transformation because it converts categorical variables into a binary vector representation, where each category becomes a separate column with a 1 or 0. This allows machine learning models to interpret categorical data without implying any ordinal relationship, which is essential for algorithms that rely on numerical distances or linear algebra.

Answer

Normalization.

Answer

Principal component analysis.

Answer

Standardization.

Question 7

Refer to the exhibit. A developer wrote a trigger to call an Einstein prediction API on lead insert. When new leads are created, the trigger fails with a 'Too many SOQL queries' error. What is the most likely cause?

Accepted Answer

The trigger is making API calls in a loop, causing governor limit issues. The trigger is making synchronous API calls to the Einstein prediction service inside a loop that iterates over each new lead. Each API call counts against the 'Number of callouts' governor limit (default 100 per transaction), and if more leads are inserted than the limit allows, the transaction fails with a 'Too many SOQL queries' error—though the error message is misleading, the root cause is exceeding the callout limit, not SOQL queries. The trigger should batch the API calls or use asynchronous processing to stay within limits.

Answer

The trigger is executed before the lead is saved

Answer

The trigger is querying a related object without an index

Answer

The Score__c field is not writeable

Question 8

Which THREE are key dimensions of data quality that directly impact AI model performance?

Accepted Answer

Consistency.. Consistency is a key dimension of data quality because AI models rely on stable patterns in the data. If the same entity is represented differently across records (e.g., 'NY' vs 'New York'), the model may learn incorrect correlations, leading to degraded prediction accuracy and unreliable outputs.

Answer

Timeliness.

Answer

Data volume.

Question 9

A company wants to use Einstein Prediction Builder to predict customer churn. Which data preparation step is essential before building the model?

Accepted Answer

Define the prediction objective and the target date field.. Einstein Prediction Builder requires you to define the prediction objective (e.g., 'Will this customer churn?') and specify the target date field that marks the event. This step is essential as it tells the model what to predict and over what time window, enabling the automated feature engineering and model training process.

Answer

Ensure the data is in a Salesforce connected data source like Data Cloud.

Answer

Create a formula field to calculate the churn probability.

Answer

Create a new custom object to store the prediction results.

Question 10

A bank uses Einstein Discovery to generate insights about loan approval decisions. After deployment, they notice the model denies loans to a higher percentage of applicants from a certain postal code. Which action should be taken to ensure responsible AI?

Accepted Answer

Audit model outcomes for fairness across demographic groups and retrain if needed. Responsible AI requires auditing model outcomes for fairness across demographic groups, even when the disparity correlates with a non-protected attribute like postal code. In Einstein Discovery, postal code can act as a proxy for protected attributes such as race or socioeconomic status, and ignoring this could lead to discriminatory lending practices. Auditing allows the team to detect and mitigate bias, and retraining with fairness constraints ensures the model aligns with ethical AI principles.

Answer

Ignore the discrepancy because postal code is not a protected attribute

Answer

Retrain the model using only recent loan data

Answer

Remove the postal code field from the model

Question 11

A company uses Salesforce Data Platform to store customer data. They want to use this data to train an AI model for lead scoring, but they are concerned about data quality. Which step should they take first to ensure the data is suitable for AI?

Accepted Answer

Profile the data to identify missing values, outliers, and inconsistencies. Profiling the data is the essential first step because it systematically identifies missing values, outliers, and inconsistencies that degrade model performance. Without this baseline assessment, any subsequent normalization or labeling would be applied to flawed data, leading to unreliable lead scoring predictions. Salesforce Data Platform supports profiling via tools like Einstein Analytics or Data Prep, which scan fields for nulls, range violations, and format errors.

Answer

Immediately normalize all numerical features

Answer

Create a labeled dataset using historical lead outcomes

Answer

Set up a data pipeline to stream data in real-time

Question 12

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 13

A company is using customer support tickets to train a model for auto-classifying issues. The dataset includes fields like 'Case Title', 'Description', 'Product', and 'Customer Name'. Which privacy concern is most critical to address before training?

Accepted Answer

Anonymize personal identifiable information (PII) in Description and Title. Anonymizing PII in the text fields is critical to avoid exposing customer information in model artifacts or predictions. Session tokens are irrelevant, and case numbers are not PII. Opt-in is a legal requirement but not directly about data preparation for AI.

Answer

Encrypt session tokens used in the support system

Answer

Remove all case numbers to prevent data leakage

Answer

Ensure all customers have opted in before using their data

Question 14

A Salesforce admin wants to use Einstein GPT to generate personalized email content for a marketing campaign. To ensure the AI does not produce responses that include sensitive customer data or violate company policies, which Salesforce feature should the admin configure?

Accepted Answer

Einstein Trust Layer. Einstein Trust Layer is the correct feature because it acts as a governance and security layer between Salesforce and the large language model (LLM). It automatically masks sensitive customer data (e.g., personally identifiable information) before the prompt is sent to the LLM and then unmasks the response, ensuring the AI never sees or exposes sensitive information. This directly addresses the admin's need to prevent responses containing sensitive data or violating company policies.

Answer

Prompt Builder

Answer

Data Cloud

Answer

Einstein Studio

Question 15

A service manager wants to automatically categorize incoming support cases based on the customer's description. Which Einstein feature should be used?

Accepted Answer

Einstein Case Classification. Einstein Case Classification uses machine learning to automatically categorize incoming support cases based on the customer's description, assigning them to predefined case fields (e.g., type, priority, product). This directly matches the requirement to automatically categorize cases from text, making it the correct choice.

Answer

Einstein Reply Recommendations

Answer

Einstein Bots

Answer

Einstein Article Recommendations

Question 16

A data analyst is troubleshooting Einstein Article Recommendations that are not showing up on the site. Which TWO checks should be performed first? (Choose 2)

Accepted Answer

Check that article view events are being captured in the data. Einstein Article Recommendations rely on user interaction data, specifically article view events, to generate personalized recommendations. If these events are not being captured, the model has no input to learn from, and recommendations will not appear. Checking event capture is a fundamental first step in troubleshooting data pipeline issues.

Answer

Ensure at least 100 articles are in the knowledge base

Answer

Confirm that article authors have the correct profile permissions

Answer

Increase the recommendation frequency from daily to hourly

Question 17

A data scientist is evaluating the performance of an Einstein Discovery model. They observe that the model has high accuracy but low precision for a specific prediction class. What does this indicate?

Accepted Answer

The model correctly predicts most instances but has many false positives for that class.. High accuracy with low precision for a specific class indicates that while the model correctly classifies the majority of instances overall, it produces a high number of false positives for that class. Precision measures the proportion of positive identifications that were actually correct, so low precision means many of the predicted positive cases are false alarms. In Einstein Discovery, this trade-off is critical when optimizing for business outcomes where false positives are costly.

Answer

The model is overfitted to the training data.

Answer

The model correctly predicts most instances but has many false negatives for that class.

Answer

The model rarely predicts that class, leading to high accuracy.

Question 18

Which TWO actions are best practices when implementing Einstein Prediction Service?

Accepted Answer

Clean the data to handle missing values and outliers.. Data quality directly impacts the accuracy and reliability of Einstein Prediction Service models. Cleaning data to handle missing values and outliers ensures that the training data is representative and reduces the risk of skewed predictions, which is a fundamental prerequisite for any machine learning model within Salesforce's AI framework.

Answer

Ignore correlated features to simplify the model.

Answer

Include all available fields in the dataset for maximum information.

Answer

Use the default field mapping without review.

Question 19

A company has a custom AI model for sentiment analysis and wants to use it in Salesforce without rebuilding. Which approach should they take?

Accepted Answer

Use Bring Your Own Model (BYOM) for Einstein. Bring Your Own Model (BYOM) for Einstein allows companies to deploy their own pre-trained AI models directly into Salesforce without rebuilding them. This approach leverages Salesforce's infrastructure for inference while keeping the custom model intact, making it the ideal solution for integrating a custom sentiment analysis model.

Answer

Use Data Export and import into external system

Answer

Use MuleSoft

Answer

Build in Apex

Question 20

A healthcare provider uses an AI system to predict patient readmission risk. The system was trained on historical data from the past five years, during which the hospital served a predominantly urban population. Recently, the hospital expanded to rural areas with different demographic and socioeconomic profiles. The AI predictions have been less accurate for rural patients, leading to misallocation of care resources. The AI Ethics committee is reviewing the system for potential bias. The model outputs a risk score from 0 to 100. The data science team has identified that the model uses features such as income, distance from hospital, and insurance type, which may correlate with race and socioeconomic status. The team wants to make the model fairer without retraining from scratch. Which approach best balances fairness and predictive accuracy?

Accepted Answer

Apply a post-processing calibration that adjusts the risk score thresholds separately for urban and rural populations to achieve equal false positive rates.. Post-processing calibration adjusts the decision thresholds for each subgroup (urban vs. rural) to equalize a fairness metric (e.g., false positive rate) without modifying the underlying model. This approach preserves the predictive signal from the original features while directly addressing the bias caused by distribution shift, making it the most practical solution when retraining from scratch is not feasible.

Answer

Remove the features income, distance, and insurance type from the model and retrain.

Answer

Continue using the current model but add a disclaimer that predictions may be less accurate for rural patients.

Answer

Retrain the model using only the latest year of data that includes rural patients.