Free CPLDA sample questions

Correct answer: Define a Contact Policy in Next-Best-Action Designer at the 'Credit Cards' Group level.

Contact Policies (suppression rules) can be defined at different levels of the taxonomy (Business Issue, Group, or Action). Defining it at the Group level ('Credit Cards') ensures that accepting an offer in that group suppresses other offers in the same group, without affecting the 'Mortgages' group.

Correct answer: Implement Context Weighting to dynamically boost the priority formula for Retention actions when Churn Risk > 0.8.

Context Weighting allows the arbitration formula to be dynamically adjusted based on customer context (like Churn Risk). This is the standard and most flexible way to re-prioritize actions based on real-time customer data without hard-coding static values.

Correct answers: Applicability Rule for the website visit check.; Contact Policy (Suppression) for the email history check.

Applicability rules determine if an action is appropriate for the customer at this time (e.g., 'Has visited website recently').

Contact Policies control the frequency of customer contact to prevent fatigue (e.g., 'Suppress if email sent in last 7 days').

Correct answer: Use the 'Final Arbitration' extension point strategy to apply conditional logic based on the Business Issue.

While NBA Designer provides a standard arbitration formula, complex variations by Business Issue often require extending the framework. The 'Final Arbitration' or specific arbitration extension points allow for custom strategy logic to modify the formula or priority based on the Issue (Sales vs Wealth Management).

Correct answer: The Suitability rule evaluated to false, which filters the action before arbitration.

In the NBA execution pipeline, Engagement Policies (Eligibility, Applicability, Suitability) are applied *before* Arbitration. If a Suitability rule returns false, the action is discarded (or effectively has 0 priority in some views) and does not participate in the final arbitration ranking.

Correct answer: Assign both treatments to the Action and configure an A/B test in the Action's 'Treatments' tab with a 20/80 split.

Pega CDH supports A/B testing directly at the Action level within the Treatments tab. You can associate multiple treatments for the same channel and define the percentage of the audience that should receive each variant.

Correct answer: Use the same Next-Best-Action strategy for both. Use a Scheduled Data Flow to write decisions to a file for the Web Portal, and the Real-Time Container service for the Mobile App.

The core principle of Center-out is using the *same* brain (Strategy/NBA configuration) for all channels. For the legacy web portal, a batch run (Scheduled Data Flow) can execute the strategy and output results to a file. For the mobile app, the Container service executes the same strategy in real-time.

Correct answer: Only properties explicitly required for Engagement Policies and Arbitration should be included to minimize memory footprint.

CIC resides in memory (often Apache Ignite/Hazelcast within Pega nodes). Loading the entire customer record (xCAR) with hundreds of unused columns wastes memory and degrades performance. Only data needed for decisioning logic (policies, arbitration, adaptive models) should be cached.

Correct answer: Decouple the data retrieval using a background Real-Time Data Flow to pre-load data into the Customer Insights Cache (xCAR).

Fetching data from slow external services during the decision request (blocking call) is a bad practice. The correct pattern is to ingest data asynchronously (e.g., via Kafka or batch) into the internal xCAR/CIC so the decision strategy reads from high-speed local cache.

Correct answer: External Data Flow with a DataSet configured for 'Database Table' utilizing the Snowflake JDBC driver.

Pega Infinity supports direct queries to external data warehouses like Snowflake via Data Sets/Data Flows, allowing strategies to access data without full ingestion (ETL). This is often referred to as federated data access or zero-copy architecture.