AI Associate Practice Test 34 — 15 Questions

Question 1

A company uses a generative AI model to create marketing copy. They want to ensure the output is accurate and not misleading. Which TWO practices should they implement?

Accepted Answer

Always have a human review the generated content before publishing. Option C is correct because human-in-the-loop review is the most reliable safeguard against inaccurate or misleading generative output; a person can catch hallucinations, unsupported claims, and brand/compliance issues before publication. Option D is correct because grounding the model with a knowledge base of verified facts (retrieval-augmented generation, RAG) supplies authoritative context and citations, which measurably reduces hallucination and keeps the copy aligned with real product and legal facts. Option A is not appropriate because raising the temperature increases randomness and creativity, which raises the risk of fabricated or misleading statements. Option B is not appropriate because disabling content filtering removes safety guardrails and can allow harmful or non-compliant text through. Option E is not appropriate because publishing unchecked model output directly is exactly the practice that permits inaccuracies and misleading claims to reach customers.

Answer

Increase the model's temperature setting to encourage creativity

Answer

Disable all content filtering to avoid blocking useful output

Answer

Use the model's output directly without checking

Question 2

You are an admin at a financial services firm. The firm wants to use Einstein Next Best Action to offer personalized product recommendations to customers on its service portal. The data includes customer profiles, transaction history, and support case history. The Einstein Next Best Action strategy is configured with a recommendation that shows a 'Savings Account' offer to customers who have a checking account. However, the recommendation is not appearing for any customers. You check the Data Flow and see that the 'Account' object data is flowing correctly. The recommendation's filter condition is: AND( Has_Checking_Account__c = true, Age__c > 18 ). You verify that many customers meet these conditions. What is the most likely reason the recommendation is not appearing?

Accepted Answer

The recommendation is not activated or published. The most likely reason the recommendation is not appearing is that it has not been activated or published. In Einstein Next Best Action, recommendations must be explicitly activated or published to become available for serving to customers; configuration alone does not make them live. Since the data flow is correct and the filter conditions are valid, the missing activation step is the typical cause of a recommendation not showing.

Answer

The 'Account' object is not supported by Einstein Next Best Action

Answer

The customer data is not being refreshed in real time

Answer

The filter condition is syntactically incorrect

Question 3

A support manager wants to automatically suggest relevant knowledge articles when agents open a case. Which Einstein feature should they enable?

Accepted Answer

Einstein Article Recommendations. Einstein Article Recommendations automatically suggests relevant knowledge articles to agents based on the case details, helping resolve cases faster.

Answer

Einstein Recommendation Builder

Answer

Einstein Next Best Action

Answer

Einstein Case Classification

Question 4

Which TWO considerations are important when labeling data for a supervised learning model?

Accepted Answer

Maintaining consistent guidelines.. Option A (Maintaining consistent guidelines) is correct because supervised learning models learn from the labels they are given, so a documented, uniformly applied labeling rubric ensures that the same input is labeled the same way across the dataset, preventing contradictory training signals that degrade model accuracy. Option B (Labeler expertise) is correct because accurate labels for domain-specific tasks (e.g., medical imaging or legal text) require annotators who understand the subject matter; unqualified labelers introduce systematic errors that the model will learn and reproduce. Option C is wrong because automated labeling is not appropriate for all tasks—it can propagate bias or errors and often needs human verification, especially for nuanced cases. Option D is wrong because inter-labeler agreement is a key quality metric; ignoring it hides ambiguity in the guidelines and reduces label reliability. Option E is wrong because labeling only a small sample limits training data volume and coverage, leading to underfitting and poor generalization.

Answer

Using automated labeling for all tasks.

Answer

Ignoring inter-labeler agreement.

Answer

Labeling only a small sample.

Question 5

According to Salesforce's AI Trust Principles, which TWO practices are essential for ethical AI deployment?

Accepted Answer

Hold the organization accountable for AI-driven outcomes.. Salesforce's AI Trust Principles emphasize organizational accountability for AI-driven outcomes, ensuring that the organization takes responsibility for the decisions and impacts of its AI systems. This principle aligns with ethical AI deployment by requiring governance, oversight, and mechanisms to address unintended consequences, rather than shifting blame to the technology itself.

Answer

Use AI to replace human jobs entirely.

Answer

Ensure models achieve 100% accuracy before deployment.

Answer

Fully automate decision-making without human review.

Question 6

A company is implementing an AI system to recommend marketing campaigns. To align with Salesforce's ethical AI principles, which practice is most important?

Accepted Answer

Ensure a human reviews AI-generated campaigns before sending. Option A is correct because Salesforce's ethical AI principles emphasize human oversight and accountability, so having a human review AI-generated campaigns before sending ensures responsible use and prevents unintended or biased outputs. This aligns with the principle of keeping humans in control of consequential decisions. Option B is incorrect because targeting only high-value customers is a business segmentation choice, not an ethical AI safeguard, and could even raise fairness concerns. Option C is incorrect because optimizing subject lines for open rates is a performance tactic, not an ethical governance practice. Option D is incorrect because removing human involvement directly contradicts the need for human oversight and accountability in ethical AI.

Answer

Target only high-value customers

Answer

Maximize open rates by optimizing subject lines

Answer

Remove human involvement to speed up campaigns

Question 7

Refer to the exhibit. An admin runs a preprocess script before training an Einstein model. Why is normalization applied to the 'AnnualRevenue' and 'NumberOfEmployees' columns?

Accepted Answer

To ensure both features contribute equally to the model. Normalization scales features like 'AnnualRevenue' and 'NumberOfEmployees' to a comparable range (e.g., 0–1 or with zero mean and unit variance). Without normalization, a feature with larger numeric values (e.g., revenue in millions) would dominate distance-based calculations in models like k-nearest neighbors or gradient descent, causing the model to undervalue the smaller-scale feature. By normalizing, both features contribute equally to the model's learning process, which is essential for many machine learning algorithms used in Einstein.

Answer

To detect outliers in the data

Answer

To remove rows with missing values

Answer

To reduce the number of features from 30 to 2

Question 8

A company uses Einstein Discovery to analyze sales data and provide recommendations. A sales rep wants to understand why a specific opportunity was predicted to close. Which Einstein feature should the rep use?

Accepted Answer

The 'Score Factors' or 'Prediction Explanation' component available on the record page.. The 'Score Factors' or 'Prediction Explanation' component on the record page is the correct answer because it provides a transparent breakdown of the top factors that influenced a specific prediction for an individual record. This feature, often called 'explainability' or 'prediction explanation', shows how each factor contributed to the predicted outcome, such as why an opportunity is likely to close. It is designed exactly for sales reps to understand the reasoning behind a prediction at the record level, enabling them to take informed actions.

Answer

The 'Model Performance' tab in Einstein Discovery Settings.

Answer

The 'Prediction Summary' report in Einstein Analytics.

Answer

The 'Data Quality' dashboard to check if the prediction is reliable.

Question 9

An administrator notices that Einstein Lead Scoring is not displaying scores for some leads. The leads have the required fields populated. What is the most likely cause?

Accepted Answer

Insufficient historical lead conversion data to train the model. Einstein Lead Scoring requires a minimum amount of historical lead conversion data to train its predictive model. If there is insufficient data, the model cannot generate scores, even if all required fields are populated. This is the most likely cause because the model relies on pattern recognition from past conversions, not just field completeness.

Answer

The lead record type is not included in the scoring model

Answer

The lead score field is not added to the page layout

Answer

The user does not have the 'View Einstein Lead Scoring' permission

Question 10

A company uses computer vision to scan receipts for expense reporting. The model performs well on high-resolution scans but poorly on blurry photos. Which improvement is most effective?

Accepted Answer

Add blurry images to the training data. Adding blurry images to the training data directly addresses the distribution mismatch between the training set (high-resolution scans) and the real-world inference inputs (blurry photos). This is data augmentation / domain adaptation, and it teaches the model to generalize to the degraded inputs it will actually see.

Answer

Decrease the learning rate

Answer

Use a larger batch size

Answer

Increase the model's number of layers

Question 11

A sales manager wants to automatically prioritize leads based on their likelihood to convert. Which Salesforce Einstein feature should be used?

Accepted Answer

Einstein Lead Scoring. Einstein Lead Scoring is the correct feature because it is specifically designed to automatically prioritize leads based on their likelihood to convert, using historical data and machine learning models. It assigns a score (0–100) to each lead, enabling sales teams to focus on high-conversion leads without manual intervention.

Answer

Einstein Opportunity Scoring

Answer

Einstein Prediction Builder

Answer

Einstein Discovery

Question 12

A service manager wants to reduce case resolution time by automatically categorizing incoming cases and suggesting relevant knowledge articles. The team has limited data science expertise. Which combination of Einstein features should be used?

Accepted Answer

Einstein Case Classification and Einstein Article Recommendations. Einstein Case Classification uses machine learning to automatically categorize incoming cases based on historical data, and Einstein Article Recommendations suggests relevant knowledge articles to agents, directly addressing the goal of reducing case resolution time without requiring extensive data science expertise.

Answer

Einstein Lead Scoring and Einstein Discovery

Answer

Einstein Bots and Einstein Conversation Insights

Answer

Einstein Vision and Language Platform

Question 13

A company wants to implement an AI-powered customer service chatbot that generates responses based on customer inquiries. To comply with GDPR requirements for automated decision-making, what must the company provide to customers?

Accepted Answer

A clear explanation of how the AI reaches its conclusions and the option to request human review.. GDPR Article 22 grants data subjects the right not to be subject to a decision based solely on automated processing that produces legal or similarly significant effects, and Article 13-15 require meaningful information about the logic involved. For an AI chatbot making decisions with significant impact, the company must provide a clear explanation of how the AI reaches conclusions and offer the option to request human review. This directly satisfies the GDPR transparency and human-intervention requirements for automated decision-making.

Answer

The ability to opt out of all AI interactions entirely.

Answer

A copy of the AI model's source code upon request.

Answer

An assurance that no personal data is used in training the AI model.

Question 14

A financial services firm uses an AI model to approve loan applications. They discover that the model denies loans at a higher rate for a certain demographic group, even when financial indicators are similar. What is the primary ethical concern?

Accepted Answer

Bias in training data. The primary ethical concern is bias in training data, which leads to the model unfairly denying loans to a demographic group even when financial indicators are similar. This is a classic example of algorithmic bias, where historical data reflects past discrimination or systemic inequalities. The model learns these biases and perpetuates them, resulting in disparate impact. Addressing this requires careful data auditing and fairness-aware machine learning techniques.

Answer

Low model accuracy

Answer

Model overfitting

Answer

Hallucination

Question 15

A data scientist is building a custom AI model using Salesforce Data Cloud to predict customer churn. They want to ensure that the model does not inadvertently use gender as a feature to avoid biased predictions. Which step is MOST appropriate?

Accepted Answer

Exclude gender from the feature set used for model training. The most direct and appropriate step to prevent gender-based bias is to exclude gender from the feature set used for model training, ensuring the model cannot learn gender-correlated patterns. This is a pre-processing fairness technique that removes the protected attribute before training. While not a complete solution (proxy variables can still encode gender), it is the standard first-line mitigation and the most appropriate answer among the options.

Answer

Use gender as a feature but ignore the model predictions for certain groups

Answer

Allow gender in training but use a post-processing technique to adjust scores

Answer

Include gender as a feature and then apply a fairness constraint during training