Pega Decision Strategy Guide

Cross-Sell on the Web Extended: Overview

• The course explains how to customize a core Next-Best-Action strategy using decision strategies.
• U+ is a retail bank using its website to enhance customer engagement and sales in real-time.
• The bank uses Pega Customer Decision Hub™ to show relevant banner ads.
• Next-Best-Action Designer helps create strategies using business rules and AI.

Extending Next-Best-Action Designer Capabilities at U+ Bank

• U+ Bank aims to meet new business needs by creating suitability rules based on customer credit scores.
• The bank requires real-time computation of credit scores based on customer profiles.
• Suitability rules will decide offer eligibility, for example only offering certain rewards to customer Barbara based on her credit score.
• U+ Bank plans to enhance its customer experience using Pega Customer Journeys for regulation and documentation.

Calculating Credit Score Using a Scorecard

• A customer’s credit score is calculated for offer suitability using a scorecard.
• A scorecard assigns importance to data pieces within a calculation.
• A scorecard uses a subset of customer property vales, which are divided into ranges and assigns scores.
• Next-Best-Action Designer implements the suitability condition using a decision strategy, referencing a Scorecard rule to decide customer credit scores.

Customer Journeys: Regulating Customer Experience

• Customer Journeys divide into stages containing multiple actions.
• Customers can advance stages by meeting entry requirements.
• Customer journeys allow for managing content and influencing best action decisions.

Decision Strategies Overview and Objectives

• Optimizes business processes by using data-driven. real-time arbitration.
• Facilitates selection the best actions for specific contexts.
• Focuses on creating customer centric interactions.
• The main objectives are:
  • Creating and configuring decision strategies using components from scratch.
  • Implementation of various components needed to build effective decision strategies.
  • Developing and validating expressions to calculate and prioritize actions.
  • Testing and refining strategies to ensure alignment with business needs.

Creating a New Decision Strategy

• Optimizes decision points in a business process or a case life cycle.
• Allows for data-driven real-time arbitration to make decisions.
• You can build a Next-Best-Label strategy to illustrate the decision strategy mechanics.
• This starts with the creation of a new strategy from scratch.
• Strategy outcomes are the actions listed in the Strategy-Results and limited to the Business issue and Group.
• The Label action selected will be the one with the lowest printing cost.
• Apply the strategy to the PegaCRM-Data-Customer class, and important customer properties become accessible.
• Use the import component when building the strategy.

Configuration and Components Selection

• A Proposition Data component is needed to define the actions in the strategy.
• The Business issue and Group are unchangeable when using an Enablement Business issue.
• The strategy imports all actions in a group by default.
• Descriptions under the component canvas will be generated automatically.

Set Property Component

• Set Property, an Enrichment component, is added to the canvas.
• The strategy aims to find the Label action with the lowest printing cost
• PrintingCost is calculated as base printing cost plus cost depending on the number of letters.
• Information in the Source components tab is populated by Proposition Data component.
• Values need to be calculated on the Target tab.
• In Pega, all inputs which begin with a dot (dot-operator) mean strategy properties are used.

Strategy Properties

• The property types are defaulted to Text, Decimal types can also be implemented.
• PrintingCost is a numeric value, so change its type to Decimal.
• Create complex calculations using the expression builder.
• PrintingCost equals BaseCost + 5 * LetterCount.
• Two types of properties are available to use in Expressions:
  • Strategy properties (Action and calculation).
  • Properties from strategy context (customer).
• Use ‘Customer dot' prefix to implement customer property into the Expression.

Prioritizing and Validating Actions

• It is best practice validate the dot-operator and test after building an Expression.
• Prioritize actions based on the lowest printing cost and add the Prioritize component.
• The prioritization could be based on an equations or existing property.
• Order how actions are presented, configured to be according to interested in lowest printing costs.
• Strategies number of returned actions can also be selected.

Testing and Refining Decision Strategies

• Testing involves checking then expanding the right-hand side test panel click save and run.
• Results can be viewed selecting a component.
• If there is more than one action it is showed as a Page.
• All values for strategy properties like Printing Cost are showed on the canvas
• The strategy is executed against a Data Transform called UseCase1.

Key points of the Demo

• Creating a decision strategy from scratch
• Configuring Proposition Data, Set Property and Prioritizing components
• Building expressions
• Exploring property categories available for expressions.
• Testing a decision strategy using a use case in a data transform.

Components and Strategy Execution

• Arrows, at runtime, carry information from the component they originate from to where they point.
• Name, BaseCost, LetterCount are available at every Select Property component.
• Each decision component Rank value is computed automatically.

Computing Printing Coat

• Printing cost totals the BaseCost and variable cost based on LetterCount.
• Filter components reference information contained inside Aggregation components.
• Apply filter conditions to the average printing cost via the dotted arrow and via component names as Expressions.
• Only actions with a 'true' condittion Expression will result in data.

Implementing Business Requirements with Decision Tables

• Business rules are implemented using a Decision Table and a set of return values.
• Actions information is received from the Filter components via the solid arrow. -A segmentation column flags actions complying with the businesss rules.

Predictive Analytics and Prioritization

• Predicitve analytics determine the propensity of actions using the Adpative Model component.
• Propensity evaluates the likelihood of interest or customer engagement with any particular action.
• The model uses and monitos customer characteristics like Age and Region.
• In the end an overall propenensitiy os calcuated for each action.
• After the Prentisy to accept values of eaach component ae set, the chain is prioritized by Prioritized component.

Prioritazion and Filtering

• A paitriozation expression calculates a priority value for each action, the higher the value, the higher the priority and value.
• In this case likelihood of acceptance weighs through propentistiy and printing cost.
• This flips Green and Black so Black is now the prinary action

Scorecard: mathematical tool.

• Scorecard is a mathematical model and it consist of of condition sets plus combiner function.
• Combiner function is for combine sum of values to produce range.
• This process is determine score of segment.
• bank had defined 3 segment. and these are high. low medium.
• Output of score card is Score and Segment Result.

Creating scorecard: main focus.

• U bank wants to build a scorecard model and customer properties.
• To build this model, Hascard, Income, and Age should be consider.
• After this process. method of combine score should be selected and also, we use Sum option.
• Score card help customer propert assign score or value.
• If customer had credict card. score is must be 100, otherwise zero.
• For this example and other calculations of result better chance of approved. means Higher result.

Building Scorecard:

• Calculating credit scores will help the bank figure out mortgage suitability.
• The bank used income and age.
• Also, if customers have a credit card, to use in the calculation (HasCards).
• Use a short scorecard description.
• The combiner is Sum .
• Using “Y” or “N” for HasCards.

Setting property Values and Ranges

• Setting numeric splits the yearly income into three ranges.
• The data model has the monthly income.

Setting Results

• After the scorecard is defined, define the results after the scorecard calculation.
• When you refresh, the scorecard calculates the max's and min's.
• There is a Result value of “Not Suitable” <250.
• Values over >250 are said to be “Suitable”.

How to test card results.

• First, go to scorecard and verify.

Demonstrations key points.

• The major part of this example was the setup of scorecards and adding property vaule

Cross Selling basics in banking

• For this situation, Credit card is shown cross selling on the web.

Decision Strategy:

• Implementing with a Decision Strategy may take place as an occurrence.
• It will sometimes affect the action at the time to get requested.

Proposition Rules:

• Proposition Rules make conditions that make most customers eligible for particular offers.
• The decision strategy that has new eligibility requirements will be implemented after the new configuration.
• The new canvas shows rules being implemented.

Parameters:

• To show the mechanics of the Engagement Strategy, one must show custom offers to the Standard card and nothing less.
• Then the strategy is tested; this makes sure the strategies output multiple variants of this type.

Eligibility & Final Configurations:

• The major goal is to ensure that Engagement has only set results for only set offers.

This guide probably details rules regarding copyright, distribution, accuracy and usage of Pegasystems software. Also discusses strategy properties, expression builder and 'Set Property' component. It contains costs calculation for product features.