CCNA Ai Capabilities Crm Questions

Why this answer

How to eliminate wrong answers

Option B is wrong because reducing the confidence threshold would cause the bot to match intents more loosely, likely increasing misclassifications rather than resolving them. Option C is wrong because disabling the 'RecentTransactions' intent would prevent the bot from handling transaction inquiries at all, which contradicts the expansion goal. Option D is wrong because creating a separate bot for transactions adds unnecessary complexity and does not fix the root cause of intent confusion; the same misclassification could occur if users are routed incorrectly.

Why this answer

Why this answer

Why this answer

How to eliminate wrong answers

Option A is wrong because Workflow Rules are a declarative automation tool for triggering actions based on record changes, not for generating AI-based content. Option C is wrong because Process Builder is a point-and-click automation tool for creating complex business processes, not for generating personalized email drafts using AI. Option D is wrong because Email Templates are static, reusable message formats that require manual selection and do not use AI to dynamically generate personalized content for each lead.

Why this answer

How to eliminate wrong answers

Option A is wrong because Flow is a declarative automation tool for building guided processes or screen flows, but it does not use AI to dynamically recommend actions based on real-time call context. Option C is wrong because Quick Actions are predefined, one-click actions (e.g., log a call, create a task) that are static and not AI-driven, so they cannot recommend the most relevant action during a call. Option D is wrong because Process Builder is a point-and-click tool for automating standard business processes (e.g., record updates, email alerts) and lacks AI capabilities to generate context-aware recommendations.

Why this answer

How to eliminate wrong answers

Option A is wrong because the error is not about enabling Einstein features globally; the admin already attempted to enable scoring, implying features are enabled. Option B is wrong because lead field mapping is not required for Einstein Lead Scoring; the feature uses standard lead fields automatically. Option D is wrong because the 'Manage Einstein' permission is not a prerequisite for enabling scoring; the admin likely already has necessary permissions if they can access the setup page.

Why this answer

How to eliminate wrong answers

Option A is wrong because Einstein Reply Recommendations suggests pre-written email responses based on context, not categorization of cases. Option C is wrong because Einstein Bots automate conversational flows and deflect cases, but they do not perform classification of case records. Option D is wrong because Einstein Article Recommendations suggests knowledge articles to agents or customers, not categorizing the case itself.

Why this answer

Why this answer

How to eliminate wrong answers

Option A is wrong because Einstein Prediction Builder is a no-code tool for creating custom predictive models on any object or field, not specifically designed for lead scoring or conversion likelihood. Option B is wrong because Einstein Relationship Health measures the strength of existing account relationships, not the conversion potential of new leads. Option D is wrong because Einstein Activity Capture automatically logs emails and events to Salesforce records, but it does not perform any predictive scoring or prioritization of leads.

Why this answer

How to eliminate wrong answers

Option A is wrong because Einstein Sentiment analyzes the emotional tone of text (e.g., positive, negative, neutral) in conversations or social posts, not for providing automated chat responses. Option B is wrong because Einstein Case Routing uses machine learning to assign cases to the best agent based on skills and availability, not to generate chat replies. Option D is wrong because Einstein Recommender suggests next-best actions or products to agents or customers based on historical data, not for automating chat responses.

Why this answer

How to eliminate wrong answers

Option A is wrong because numeric prediction is used for forecasting a continuous numerical value (e.g., revenue amount), not a categorical outcome like churn. Option B is wrong because multi-class classification predicts among three or more categories (e.g., product type), whereas churn is a binary yes/no outcome. Option C is wrong because regression is a type of numeric prediction for continuous values, not suitable for a binary classification task.

Why this answer

Why this answer

Why this answer

How to eliminate wrong answers

Option A is wrong because Data Export is a tool for exporting Salesforce data to external systems, not for importing or running custom AI models within Salesforce. Option C is wrong because MuleSoft is an integration platform for connecting applications and data, not a service for deploying custom AI models into Salesforce's AI framework. Option D is wrong because building the model in Apex would require rewriting the entire model from scratch in Apex code, which is impractical for complex machine learning models and defeats the purpose of using an existing custom model.

Why this answer

How to eliminate wrong answers

Option A is wrong because Email Templates are used for standardizing email communications, not for generating contract clauses with preapproved language. Option B is wrong because Flow is a process automation tool for orchestrating actions and approvals, not a content generation system for legal clauses. Option C is wrong because Document Templates provide static document structures but lack the AI-driven clause selection and generation capabilities needed for dynamic, preapproved clause insertion.

Why this answer

How to eliminate wrong answers

Option A is wrong because Queues are used for routing work items (like cases or leads) to a group of users based on assignment rules, not for building conversational AI or handling customer inquiries directly. Option C is wrong because Omni-Channel is a routing engine that distributes work across channels (chat, phone, etc.) to available agents, but it does not provide AI-driven chatbot capabilities or automated responses. Option D is wrong because Case Assignment Rules automatically assign cases to users or queues based on criteria, but they lack the conversational AI and natural language understanding needed to interact with customers and answer order status questions.

Why this answer

Why this answer

How to eliminate wrong answers

Option A is wrong because resetting and retraining the model from scratch would discard all historical learning and does not address the missing real-time event data; it is an overreaction that would degrade performance temporarily. Option B is wrong because an Einstein Bot is designed for conversational interactions, not for passively capturing behavioral events like add-to-cart; it would add unnecessary complexity and user friction without solving the data integration gap. Option C is wrong because disabling recommendations for email and ads does not affect the capture of real-time events on the website; the issue is about data ingestion, not channel distribution.

Why this answer

How to eliminate wrong answers

Option A is wrong because retraining the model with only two predictors removes other potentially valuable influences and violates the principle of using the model as generated by Einstein Discovery, which automatically selects the most predictive features. Option B is wrong because deleting an influence from the story does not change the underlying model; it only hides the field from the UI, and the prediction still uses 'LastDonationAmount' internally, so this does not make the output more actionable. Option C is wrong because Einstein Discovery does not allow manual adjustment of influence weights in story settings; the influence percentages are determined by the model's algorithm and cannot be overridden by an administrator.

Why this answer

Why this answer

Why this answer

Why this answer

Why this answer

How to eliminate wrong answers

Option B is wrong because if Einstein features were not enabled, no Einstein functionality would work at all, not just scoring for some accounts. Option C is wrong because the 'View Scores' permission controls user visibility of scores, not the generation of scores by the AI engine. Option D is wrong because Einstein Account Scoring uses standard fields and does not require custom fields; it leverages existing data like industry, revenue, and engagement history.

Why this answer

How to eliminate wrong answers

Option A is wrong because the JSON shows lead scoring is enabled (the 'leadScoring' object has 'status' set to 'Enabled'), so lead scoring is not disabled for the org. Option C is wrong because the JSON does not contain any field indicating a retraining schedule for the lead scoring model; the retraining frequency is not specified in this configuration snippet. Option D is wrong because the JSON includes a 'scoringFields' array under 'leadScoring' that lists specific fields like 'Industry' and 'AnnualRevenue', indicating the admin has configured scoring fields for leads.

Why this answer

How to eliminate wrong answers

Option A is wrong because opportunity team members are not a prerequisite for Einstein Opportunity Scoring; the model scores opportunities based on field values and activities, not team assignments. Option B is wrong because Einstein Opportunity Scoring is available on Professional Edition with the required add-on licenses, so edition alone does not prevent scoring. Option D is wrong because the admin already confirmed all required fields are present, and missing custom fields would not block scoring unless they are specifically mapped as required by the model, which is not the case here.

Why this answer

How to eliminate wrong answers

Option A is wrong because changing the outcome to 'Escalated' alters the target variable entirely, which would require retraining a new model for a different prediction task and does not address the accuracy of the original case closure model. Option C is wrong because removing the 'Subject' field may discard valuable textual features that could help predict closure time; unless the field is irrelevant or noisy, reducing features often harms model performance. Option D is wrong because adding more fields like 'Comments' without ensuring data quality or relevance can introduce noise and increase dimensionality, potentially degrading accuracy rather than improving it.

Why this answer

Why this answer

How to eliminate wrong answers

Option A is wrong because Einstein Article Recommendations suggests knowledge base articles for agents or customers to read, not actionable next steps during a live chat. Option B is wrong because Einstein Bots are automated chatbots that handle customer interactions without a human agent, not a feature that recommends actions to a live agent. Option C is wrong because Einstein Reply Recommendations suggests pre-written text responses for agents to send in chat, which is a subset of communication assistance, not a broader recommendation of business actions like transfers or promotions.

Why this answer

How to eliminate wrong answers

Option A is wrong because retraining the model is not the first step; the issue is likely a lack of sufficient training data, not a model malfunction. Option B is wrong because the 'View Einstein Scores' permission controls visibility of scores, not the generation of scores; scores are generated regardless of user permissions. Option D is wrong because waiting 48 hours is unnecessary; the model updates daily, but the root cause is insufficient historical data, not a delay in processing.

Why this answer

How to eliminate wrong answers

Option A is wrong because increasing the maximum score does not address the underlying data mismatch; it merely scales the output range without improving the model's ability to distinguish converters from non-converters. Option C is wrong because disabling field-level security for scoring fields would expose sensitive data and violate Salesforce security best practices, and it does not affect the model's training data or prediction logic. Option D is wrong because lowering the lead conversion threshold artificially inflates conversion rates without fixing the model's accuracy; it treats the symptom (low conversion) rather than the cause (stale training data).

Why this answer

Why this answer

How to eliminate wrong answers

Option A is wrong because Custom Reports are used for creating ad-hoc data views and analyzing past performance, not for generating predictive insights about future purchase likelihood. Option B is wrong because List Views are static filters that display records based on predefined criteria, lacking any machine learning or predictive capability. Option C is wrong because Campaigns are used to manage marketing outreach and track engagement, not to predict which customer segments will purchase a product.

Why this answer

How to eliminate wrong answers

Option B is wrong because the question explicitly states both queries include the custom field Score__c, and there is no mention of a date function in either query; the performance difference is due to the filter on Score__c, not a date function. Option C is wrong because Score__c is a custom field used by Einstein Lead Scoring and is not automatically indexed; if it were indexed, the second query would likely be faster, not slower. Option D is wrong because Einstein Lead Scoring jobs run asynchronously in the background and do not directly impact the performance of individual SOQL queries; the slowdown is due to query selectivity, not concurrent job execution.

Why this answer

How to eliminate wrong answers

Option A is wrong because predicting a numeric value (regression) would produce a JSON with a single numeric 'value' field and no categorical 'predictedValue' or 'probability' fields, whereas the exhibit shows a 'predictedValue' with a string label. Option B is wrong because predicting a text field (e.g., a free-form string) would require a different model type (like Einstein Text Prediction) and the JSON would contain a 'text' or 'value' field, not a binary 'predictedValue' with two possible states. Option C is wrong because the JSON does not include any retraining status indicators such as 'retrainingInProgress: true' or 'modelStatus: retraining'; the model is deployed and serving predictions, as shown by the presence of a score.

Why this answer

Why this answer

Why this answer

Why this answer

How to eliminate wrong answers

Option A is wrong because increasing sync frequency does not address the root cause of emails being filtered out by domain; it only changes how often the system checks for new data. Option C is wrong because manually logging emails bypasses the automation that Einstein Activity Capture is designed to provide, and it is not a scalable solution for ongoing capture. Option D is wrong because the 'Einstein Activity Capture User' permission set is assigned to users, not to email domains; domains cannot be assigned permission sets.

Why this answer

How to eliminate wrong answers

Option B is wrong because increasing the data retention period does not fix the immediate issue of stale or incomplete data; it only keeps older records longer without addressing data quality or refresh delays. Option C is wrong because rebuilding the model with different fields does not resolve the underlying data refresh problem; the model will still train on incomplete or outdated data, leading to inconsistent scores. Option D is wrong because increasing training frequency to weekly does not help if the data itself is not being refreshed; the model would simply train more often on the same stale data, perpetuating inaccuracies.

Why this answer

How to eliminate wrong answers

Option A is wrong because the timing of trigger execution (before vs. after save) does not cause a 'Too many SOQL queries' error; that error is a governor limit issue, not a save-order issue. Option B is wrong because querying a related object without an index would cause a 'Too many rows' or performance warning, not a 'Too many SOQL queries' error—the error specifically indicates the number of SOQL queries (or callouts) exceeded the limit, not a missing index. Option D is wrong because the Score__c field being non-writeable would produce a field-level validation or DML exception, not a governor limit error like 'Too many SOQL queries'.

Why this answer

Why this answer

How to eliminate wrong answers

Option A is wrong because Chatter does not need to be disabled for Einstein Activity Capture to work; in fact, Chatter can remain enabled and is unrelated to the email/calendar sync process. Option B is wrong because Sales Cloud Einstein licenses are not a prerequisite for Einstein Activity Capture; the feature is available with standard Salesforce licenses and does not require an add-on Einstein license. Option C is wrong because Einstein Activity Capture is not automatically enabled when email integration is configured; it requires explicit setup, including OAuth authorization from each user, and is not a default consequence of email integration.