01. A bank wants each customer to see the action that is both relevant to them and valuable to the business, chosen in real time from the actions they qualify for.
Which Customer Decision Hub mechanism produces this selection?
a) The engagement policy, which filters actions for eligibility only
b) The adaptive model, which outputs only a propensity score
c) The business hierarchy, which groups actions by issue and group
d) Arbitration
02. A claims case type needs an automated decision at a specific point in its flow so that low-risk claims proceed automatically while high-risk claims are routed for review.
Which two statements about using Pega Process AI in this case type are correct?
(Choose two.)
a) Process AI requires the decision to be made by an outbound campaign rather than inside the case flow
b) The prediction's outcome can drive routing so the case follows a different path depending on the predicted risk
c) Case-level predictions are limited to reporting and cannot influence the flow's path
d) The prediction can only be evaluated after the case is resolved and closed
03. A new team member confuses topic classification with entity extraction.
Which two statements correctly distinguish these capabilities?
(Choose two.)
a) Entity extraction pulls specific structured values out of the text
b) Topic classification returns positive, neutral, or negative tone
c) Topic classification identifies what the whole message is about
d) Entity extraction identifies which language the message is written in
04. A service organization finds that some cases breach their Service-Level Agreement before staff realize the cases are at risk. The team wants Process AI to help agents act earlier on the cases most likely to breach.
Which prediction best supports this goal?
a) A prediction that estimates the risk of a case missing its SLA, so at-risk cases can be prioritized or escalated
b) A prediction of the customer's propensity to accept a cross-sell offer during the interaction
c) A prediction that flags whether a transaction is fraudulent and should be held for review
d) A prediction of the sentiment expressed in the customer's inbound email
05. An AI governance review at a bank must address transparency, fairness, and change control.
Which three governance mechanisms in Pega map to these concerns?
(Choose three.)
a) Revision Management approval workflows with an audit trail of changes
b) Model transparency scores together with per-issue transparency thresholds
c) Adaptive model self-learning that raises propensity over time
d) Ethical bias policies that test predictions against sensitive predictors
06. An organization is deciding between Pega Process AI and Pega Customer Decision Hub for a new requirement.
Which requirement is best served by Pega Process AI rather than Customer Decision Hub?
a) Arbitrating one-to-one next-best-action offers across a customer's inbound and outbound channels
b) Embedding real-time predictions and decisions inside a running case or business process
c) Managing an always-on customer profile that drives personalized marketing engagement
d) Scheduling outbound campaigns that promote cross-sell and up-sell actions
07. Before arbitration ranks actions, engagement policies decide which actions a customer qualifies for.
Which two purposes do engagement policies serve in Next-Best-Action?
(Choose two.)
a) Determining eligibility — which customers can receive a given action
b) Applying suitability and applicability rules so only appropriate actions proceed to arbitration
c) Assigning the transparency rating of the underlying model
d) Selecting the single highest-priority action to present
08. In a payments case type, a Process AI fraud prediction returns a low fraud risk for a transaction.
What is the typical benefit of this result for the process?
a) The case is automatically held and routed to a fraud specialist for inspection
b) The customer is presented with the next-best-action marketing offer for the interaction
c) The case's SLA timer is reset to allow additional time for review
d) The case can be processed straight through without manual review, improving efficiency
09. U+ Bank routes incoming emails to the correct department and also wants to pull out the customer's account number, which always follows the pattern of two letters followed by ten digits.
Which entity-extraction approach best fits this requirement?
a) Model-based extraction trained on a large labeled corpus
b) Sentiment analysis of the email body
c) Keyword or rule-based extraction using a regular expression that matches the pattern
d) Language detection on the email subject line
10. Which two operational outcomes can Pega Process AI predictions directly support within case management?
(Choose two.)
a) Scheduling the outbound wave that sends retention offers to churn-risk customers
b) Holding transactions that are likely to be fraudulent so they receive closer review before proceeding
c) Selecting the single best marketing offer to present on the bank's website
d) Flagging cases at risk of missing their SLA so they can be escalated