Fraud Alert Suppressed by Global Contact Policy
A bank uses Pega CDH for inbound and outbound NBA across 5.5 million customers. A new fraud-awareness action is intended to appear in online banking when a customer has a recent card-not-present transaction over $500. The action is in the Servicing issue, Card Security group. Contact policy caps all outbound and inbound offers at three per 14 days and suppresses any customer who received a fraud alert in the last 30 days. The fraud team reports that genuine fraud victims are not seeing the awareness message because they already received a transaction-decline SMS, which is modeled as a separate action and consumes the contact policy. The strategy also uses a customer-complaint-in-last-7-days suppression that was intended for marketing but is applied globally. Customer service wants the fraud message to always appear when the condition is true, even if it exceeds the cap.
1. How would you separate mandatory servicing actions from marketing actions in the NBA design?
Read the worked answer
The root cause is a global contact policy that treats a fraud SMS and a marketing offer as equivalent. Define separate contact categories: service and marketing. Service actions get their own cap or no cap, and marketing actions keep the three-per-14-days cap. In the NBA design, put the fraud action in a dedicated service issue with higher priority, and if the product supports mandatory actions, mark it mandatory. Otherwise, implement a pre-arbitration service check that returns the fraud message before marketing arbitration.
Also remove the global suppression that blocks customers who received a fraud alert. A service message about fraud should not suppress another service message about fraud unless there is a deliberate frequency rule. Test with a customer who has both a decline SMS and a card-not-present transaction.
2. What is wrong with applying the complaint suppression globally, and how should it be scoped?
Read the worked answer
The complaint suppression is a marketing suitability rule, but applying it globally also suppresses service messages. That is why a customer with a recent complaint may not see a fraud warning. Scope the rule by action category or issue so it applies only to marketing actions. Use the engagement policy's applies-to setting or a condition on the action category.
Verify with a test customer who has a recent complaint and a fraud condition: the fraud message should still appear, while marketing offers are suppressed. Document the intent of the rule so future changes do not re-globalize it.
3. The fraud team wants a guarantee. What design pattern would you use?
Read the worked answer
Use a mandatory service pattern: a dedicated service issue with its own contact-policy exception and a pre-arbitration check that returns the fraud message when the trigger is true. Store the trigger as decision data from the real-time transaction event, and add a frequency rule such as once per day plus an expiry. This makes the message reliably delivered without letting it become unlimited spam.
Define the guarantee precisely: considered and delivered unless a higher-priority compliance block applies. Monitor trigger-to-impression rate and false positives. If the platform version does not support mandatory flags, the pre-arbitration strategy is the safer implementation.