Creating an Inline Service in the Oracle Real-Time Decisions (RTD) Environment

Purpose

This tutorial shows you how to configure, deploy, analyze, and update an Inline Service using the components of the Oracle Real-Time Decisions ( Oracle RTD) environment.

Time to Complete

Approximately 2 hours

Topics

This tutorial covers the following topics:

 Overview
 Scenario
 Prerequisites
 Creating a New Inline Service
 Creating Models for Call Analysis
 Generating Offers Based on Performance Goals
 Configuring Inline Service to Learn on Offer Acceptances
 Using the Model to Influence Offer Generation
 Summary
 Related information

Viewing Screenshots

 Place the cursor over this icon to load and view all the screenshots for this tutorial. (Caution: This action loads all screenshots simultaneously, so response time may be slow depending on your Internet connection.)

Note: Alternatively, you can place the cursor over each individual icon in the following steps to load and view only the screenshot associated with that step.

Overview

The Oracle Real-Time Decisions (Oracle RTD) platform allows you to develop enterprise software solutions that analyze business process behavior and make recommendations in real-time, allowing you to identify and address problems and opportunities as soon as they emerge. In this tutorial you learn how to configure, deploy, analyze, and update an Inline Service. Inline Service refers to the configured application that is deployed using the components in the RTD environment. Oracle RTD consists of five components: Decision Studio, Real-Time Decision Server, Decision Center, Administration (JMX), and Load Generator. Inline Services are configured and deployed using Decision Studio and analyzed and updated using Decision Center. Inline Services run on Real-Time Decision Server. An Inline Service can gather data and analyze characteristics of enterprise business processes on a real-time and continuous basis. It also leverages that data and analysis to provide decision-making capability and feedback to key business processes.

Back to Topic List

Scenario

The Inline Service you build in this tutorial is based around a fictitious credit card company's call center. The Inline Service collects data about the customer and the call center operational system, and then analyzes information about the call and its resolution. The goal of this Inline Service is to analyze the patterns about calls, reasons for calling, and customers. You begin by building a basic Inline Service. In later sections, you extend the capability of this Inline Service to provide recommendations to the CRM system on cross selling offers and then to add feedback to the service on the success of its recommendations.

Back to Topic List

Prerequisites

Before starting this tutorial, you should:

1.

Have access to or have installed Oracle Real-Time Decisions (RTD), including populating the CrossSell example data provided with Oracle RTD. For more information, please refer to the administration documentation titled Installation and Administration of Oracle RTD ( section 2.3, if using RTD 2.2)

 

2.

Set up a user named sdsu using JMX.

 

3.

Download the Tutorial Code.doc file, which provides scripting solutions for the tutorial.

 

Back to Topic List

Creating a New Inline Service

The goal of this topic is to create a new Inline Service. You have finalized the business requirements for your new RTD project and you are now creating a basic Inline Service that meets those project requirements. This Inline Service retrieves customer information from a data source using a minimum number of objects, including data source, entity, and Informant. You are going to add more features to this inline service in later topics.

To create a new Inline Service, you perform the following steps:

 Create an Inline Service project
 Configure the Application element
 Create a data source and import data fields
 Create and map entities
 Deploy and test the Inline Service

Back to Topic List

To create a new Inline Service project, perform the following steps:

1.

To access Decision Studio, go to the client computer where you installed the Oracle RTD client tools and run RTD_HOME\eclipse\eclipse.exe where RTD_HOME is the Oracle RTD install directory. For example, C:\OracleBI\RTD\.

 

2.

Select File > New > Inline Service Project.

 

3.

Enter Tutorial for the project name.

 

4.

Select the Basic Template from the metadata template list.

 

5.

Verify that the Create new project in workspace option is selected.

 

6.

Click Finish.

 

Back to Topic

To configure the application element, perform the following steps:

1.

In the Inline Service Explorer, expand the Tutorial > Service Metadata folder.

 

2.

Double-click the Application element.

 

3.

Enter This is a tutorial application in the Description field.

 

4.

Click the Permissions tab.

 

5.

Click the Add button. This opens the Add User or Group window.

 

6.

Click Select Server. This opens the Connect to a server window.

 

7.

Enter sdsu for the User name and password and check Save password.

 

8.

Click Connect. Upon connection, the Connect to a server window is closed.

 

9.

In the Add User or Group window, enter sdsu in the User and Group field.

 

10.

Click OK.

 

11.

In the Permissions section, grant all privileges by clicking the Granted field for each row. Granting the Deploy Service from Studio option automatically checks Open Service and Open Service for Reading options.

 

Back to Topic

To create a data source and import data fields, perform the following steps:

1.

Select the Data Sources element in the Inline Service Explorer.

 

2.

Right-click the Data Sources element and select New SQL Data Source.

 

3.

Enter Customer Data Source in the Display Label field.

 

4.

Click OK. This opens the Customer Data Source window in the editor area.

 

5.

Click the Import button.

 

6.

In the Import Database Table window, examine and accept the RTD server parameters.

 

7.

Click Next.

 

8.

Select the SDDS data source from the JDBC Data Source list.

 

9.

Select the CROSSSELLCUSTOMERS table from the list.

 

10.

Click Finish.

 

11.

In the Output list on the left, select the ID column.

 

12.

Click the arrow pointing to the right to add this column to the Input list.

 

13.

Retain only the following seven output attributes in the Output list. Use the Remove command in the Output list to delete other attributes.
Age
HasCreditProtection
Language
LastStatementBalance
MaritalStatus
NumberOfChildren
Occupation

 

14.

Select File > Save All to save the metadata.

 

Back to Topic

To create and map entities, perform the following steps:

 Create an entity for customer attributes
 Create a session entity for the customer
 Map entities

To create an entity for customer attributes, perform the following steps:

1.

Select the Entities element in the Inline Service Explorer.

 

2.

Right-click the Entities element and select New Entity.

 

3.

Enter Customer in the Display Label field.

 

4.

Click OK.

 

5.

Click the Import command button in the editor area for the Customer entity.

 

6.

In the Data Source Selection window, select Customer Data Source.

 

7.

Click OK.

 

8.

Set the default value for each attribute as shown. Since Decision Studio encloses the string values with double quotes automatically, do not include them when you set default values.

Age = 35

HasCreditProtection = No

Language = English

LastStatementBalance = 1000

MaritalStatus = Single

NumberOfChildren = 0

Occupation = Student

 

9.

Click the Add Key command button.

 

10.

In the Key Name window, enter customerId in the Display Label field.

 

11.

Select Integer for the data type from the drop-down list.

 

12.

Click OK.

 

13.

Select File > Save All to save the Tutorial project. You will see several errors in the Problems View. This is expected because the mapping definition of the Customer entity attributes to its data source is incomplete. Proceed to the next section in order to complete the mapping definition.

 

Back to Topic

To create a session entity for the customer entity, perform the following steps:

1.

Double-click the Session entity under the Entities element folder.

 

2.

Click the Add Attribute command button in the Session editor area.

 

3.

Enter customer in the Display Label.

 

4.

Change the data type for customer attribute to Other... in the drop-down list.

 

5.

Select the Customer entity under the Entity Types folder.

 

6.

Click OK.

 

7.

Click OK.

 

8.

Click the Select command button to add a session key.

 

9.

In the Select entity keys window, expand Session > customer and check customerId as the session key.

 

10.

Click OK.

 

11.

Verify that the customerId session key is added to the session.

 

12.

Select File > Save All to save the project.

 

Back to Topic

To map the customer entity with the customer session entity, perform the following steps:

1.

Double-click the Customer entity under the Entities element folder.

 

2.

Click the Mapping tab of the Customer entity object.

 

3.

In the Data Source Input Values section, click the Customer Data Source record.

 

4.

Click the ellipsis button (button with three dots) in the Input Value field.

 

5.

In the Edit Value window, select the Attribute or Variable option for the value source, expand the tree to Session > customer, and select the customerId value.

 

6.

Click OK.

 

7.

Select File > Save All to save the project.

 

Back to Topic

To deploy and test an Inline Service, perform the following steps:

 Create an Informant for testing the Inline Service

Deploy the Inline Service

 Test the Inline Service

To create an Informant for testing the Inline Service, perform the following steps:

1.

Expand the tree in the Inline Service Explorer view to Tutorial > Service Metadata > Integration Points > Informants.

 

2.

Right-click the Informants object and select New Informant.

 

3.

Enter Testing in the Display Label field.

 

4.

Click OK.

 

5.

Click Advanced.

 

6.

In the Edit advanced properties window, uncheck the Show in Decision Center option.

 

7.

Click OK.

 

8.

Click Select.

 

9.

Select customerId as the session key.

 

10.

Click OK.

 

11.

Click the Logic tab.

 

12.

Either enter the code below or copy and paste the contents from the Lab 7-A section in the Tutorial Code.doc document.

logInfo("Customer age = " + session().getCustomer().getAge());

 

13.

Select File > Save All.

 

Back to Topic

To deploy the Inline Service, perform the following steps:

1.

Select Project > Deploy.

 

2.

In the Deploy Inline Service window, select the Tutorial project from the Project drop-down list.

 

3.

Verify that the deployment state is set to Development.

 

4.

Verify that the server is set to localhost:8080.

 

5.

Check the following options: Release Inline Service lock and Terminate active sessions (used for testing).

 

6.

Click Deploy. It takes few seconds for the deployment to complete. A message appears on the message tab at the lower left indicating that the Inline Service is deployed successfully.

 

Back to Topic

To test the Inline Service, perform the following steps:

1.

Select Window > Show View > Test to open the test window view in Testing Informant editor area.

 

2.

Verify that the Testing Informant is selected in the Integration Point drop-down list.

 

3.

Enter 1 for the value of the customerId field.

 

4.

Click the Execute Request button (green play button).

 

5.

Click the Log tab.

 

6.

Verify that the customer's age is returned.

 

7.

Test the Informant with several more values for customerId and monitor the logs.

Back to Topic

Back to Topic List

Creating Models for Call Analysis

To create models for call analysis, you perform the following steps:

 Create Informants and external systems

Create choice groups and models for call analysis

 Populate a call model using the Load Generator utility

Create Informants and External Systems

The goal of this topic is to create Informants and associate them with customer sessions and external systems. You have built a new Inline Service and tested its basic functionality. You are adding Informants and associating them with external systems that Informants use to send requests. You are also going to configure Informants to receive session key information in order for them to gather and process data based on the session.

To create Informants and external systems, perform the following steps:

 Create a Call entity

Associate Call and Session entities

 Create external systems
 Create Informants
 Test and debug Informants

To create a new Call entity, perform the following steps:

1.

In the Inline Service Explorer, expand the tree to Tutorial > Service Metadata > Entities.

 

2.

Right-click Entities and select New Entity.

 

3.

Enter Call in the Display Label field.

 

4.

Click OK. This opens the Call window in the editor area.

 

5.

In the Definition tab, click Add Attribute.

 

6.

Enter agent in the Display Label field and verify that String is selected in the Data Type drop-down list.

 

7.

Click OK.

 

8.

Add two more attributes with the following properties:

Display Label Data Type
length Integer
reason code Integer

 

9.

Verify your work looks like the screenshot:

 

10.

Select File > Save All.

 

Back to Topic

To associate Call and Session entities, perform the following steps:

1.

Double-click the Session entity under Entities folder.

 

2.

Click Add Attribute in the Session window.

 

3.

Enter call in the Display Label field and select Other... from the Data Type list. The Select a Type window opens.

 

4.

Expand Entity Types and select the Call entity.

 

5.

Click OK.

 

6.

Click OK.

 

7.

Select File > Save All.

 

Back to Topic

To create external systems, perform the following steps:

1.

In the Inline Service Explorer, expand the tree to Integration Points > External Systems.

 

2.

Right-click External Systems and select New External System.

 

3.

Enter IVR in the Display Label field.

 

4.

Click OK.

 

5.

Right-click External Systems and select New External System.

 

6.

Enter CRM in the Display Label field.

 

7.

Click OK.

 

Back to Topic

To create new Informants, perform the following steps:

1.

In the Inline Service Explorer, in the Integration Points folder, right-click the Informants folder and then select New Informant.

 

2.

Enter Call Begin in the Display Label field.

 

3.

Click OK.

 

4.

Click Select in Call Begin editor area.

 

5.

Select customer / customerId as the session key.

 

6.

Click OK.

 

7.

Set the External System to IVR in the drop-down list and enter 1 in the Order field.

 

8.

Click the Logic tab.

 

9.

Copy the contents from the Lab 8-A section of the Tutorial Code.doc document and paste it in the Logic field.

 

10.

Select File > Save All.

 

11.

Right-click the Informants folder and then select New Informant.

 

12.

Enter Service Complete in the Display Label field.

 

13.

Click OK.

 

14.

Click Select in Service Complete editor area.

 

15.

Select customer / customerId as the session key.

 

16.

Click OK.

 

17.

Set the External System to CRM in the drop-down list and enter 2 in the Order field.

 

18.

Click the Add button to add an incoming parameter.

 

19.

Enter agent in the Display Label field and verify that String is selected in the Data Type field.

 

20.

Click OK.

 

21.

Add two more incoming parameters with the following properties.

Display Label Data Type
length Integer
reason code Integer

Verify your work:

 

22.

Click the Session Attribute field for the agent parameter.

 

23.

Click the ellipsis ( ...) button.

 

24.

In the Request Attribute Assignment window, expand the tree and select Session > call > agent.

 

25.

Click OK.

 

26.

Map the other two incoming parameters with the following session attributes. 

length call / length
reason code call / reason code

 

27.

Select File > Save All.

 

28.

Right-click the Informants folder and select New Informant.

 

29.

Enter Call End in the Display Label field.

 

30.

Click OK.

 

31.

Click Select in the Call End editor area.

 

32.

Select customer / customerId as the session key.

 

33.

Click OK.

 

34.

Set the External System to CRM in the drop-down list, enter 5 in the Order field, and check Force session close.

 

35.

Click the Logic tab.

 

36.

Copy the contents from the Lab 8-B section of the Tutorial Code.doc document and paste it in the Logic field.

 

Back to Topic

To test and debug the new Informants, perform the following steps:

1.

Select File > Save All.

 

2.

Select Project > Deploy.

 

3.

In the Deploy Inline Service window, ensure that the Tutorial project is selected from the drop-down list.

 

4.

Click Deploy. This may take a few moments. Verify that a message appears on the message tab at the lower left indicating that the Inline Service is deployed successfully.

 

5.

In the Test view, select Call Begin as the integration point from the drop-down list and enter 1 for customerId.

 

6.

Click the Execute Request button.

 

7.

Click the Log tab.

 

8.

Verify that the Call Begin Informant created a session for this customerId.

 

9.

Select the Call End integration point from the drop-down list.

 

10.

Click the Execute Request button again.

 

11.

Verify that the Call End Informant recorded the session end point.

 

12.

Test these two Informants with several more values for customerId and monitor the logs.

 

Back to Topic

To create choice groups and models for call analysis, perform the following steps:

 Create a choice group

Create a choice model

 Test the model

To create a choice group, perform the following steps:

1.

In the Inline Services Explorer, expand the tree to Tutorial > Service Metadata > Choices.

 

2.

Right-click Choices and select New Choice Group.

 

3.

Enter Call Reason in the Display Label field.

 

4.

Click OK. This opens the Call Reason window in the editor area.

 

5.

In the Choice Attributes tab, click Add.

 

6.

Enter code in the Display Label field and select Integer from the Data Type drop-down list.

 

7.

Click OK.

 

8.

Verify that this choice attribute is added to the Call Reason choice group. Choice attributes defined on a choice group are available to lower level choice groups and choices.

 

9.

In the Inline Service Explorer, select the Call Reason choice group under the Choices folder.

 

10.

Right-click the Call Reason choice group and select New Choice. Be sure to pick New Choice, not New Choice Group.

 

11.

Enter Check Balance in the Display Label field.

 

12.

Click OK. This opens the Check Balance window in the editor area.

 

13.

In the Attribute Values tab, set the code Attribute Value to 17. You can either enter the value directly on the Attribute Value column or click the ellipsis command on this column; this opens a window where you can set this constant value.

 

14.

Add three more choices under the Call Reason choice group using the information from the following table:

Choice Code
Make Payment 18
Rate Inquiry 19
Other 20

 

15.

Select File > Save All.

 

Back to Topic

T o create a choice model, perform the following steps:

1.

Select the Models element under Tutorial > Service Metadata.

 

2.

Right-click the Models folder and select New Choice Model.

 

3.

Enter Reason Analysis in the Display Label field.

 

4.

Click OK. This opens the Reason Analysis window in the editor area.

 

5.

Uncheck the Use for prediction option.

 

6.

In the Choice tab, use the drop-down list to set the Choice Group target to Call Reason.

 

7.

Replace the existing value in the Label for Choice field by typing Reason for call.

 

8.

Click the Learn Location tab.

 

9.

Select the On Integration Point option.

 

10.

Click Select.

 

11.

Select Service Complete in the Integration Points list.

 

12.

Click OK.

 

13.

Double-click the Service Complete Informant under the Integration Points > Informants folder.

 

14.

Click the Logic tab.

 

15.

Copy the contents from the Lab 8-C section of the Tutorial Code.doc document and paste it in the Logic field. The logic tells the model which Reason Code was selected for the call.

 

16.

Select File > Save All.

 

Back to Topic

To test the choice model, perform the following steps:

1.

Select Project > Deploy and then click the Deploy button in the Deploy Inline Service window to redeploy the Tutorial project into the RTD server.

 

2.

Verify that a message appears on the message tab at the lower left indicating that the Inline Service is deployed successfully.

 

3.

In the Test view, select the Service Complete Informant in the Integration Point drop-down list and enter the following values:

customerId 7
agent John
length 21
reasonCode 18

 

4.

Click the Execute Request button.

 

5.

If necessary, click the Log tab.

 

6.

Verify that the Service Complete Informant added the Make Payment choice to the model. This verifies that your Service Complete Informant populates the model you created for analyzing reasons for calls.

 

7.

Test this Informant with several more values for customerId, agent, length, and reason code variables and monitor the logs to see that these call reasons are added to the model.

 

Back to Topic

To populate a call model using the Load Generator utility, perform the following steps:

 Define the Load Generator script

Add variables to the Load Generator script

 Add integration point calls to the script
 Run the script and view simulated call reason analysis in Decision Center

To define the Load Generator script using the Load Generator utility, perform the following steps:

1.

To open the Load Generator Utility, go to the client computer where you installed Oracle RTD and run RTD_HOME\scripts\loadgen.cmd .

 

2.

Click Create a new Load Generator script.

 

3.

Click the General tab.

 

4.

Verify that the Client Configuration File is set to RTD_HOME\client\clientHttpEndPoints.properties . If not, click the ellipsis command and navigate to this folder location to select this file. Please note: If the RTD server is not on the same computer as Load Generator, or RTD was not set to port 8080, then you need to modify this file ( clientHttpEndPoints.properties). There is a line at the end of the file that identifies the host name and IP, which by default it is HTTP1.url = http://localhost:8080. Update this value if your RTD server is on a different computer.

 

5.

Enter Tutorial for the Inline Service. Verify that Think Time is set to the Fixed Global Think Time option.

 

6.

Set the Think Time to 0. Setting Think Time to 0 increases the throughput while reducing the simulation load time. The reason 0 increases throughput is because 0 is the number of seconds between when requests (1 request = 1 integration point call) are sent to the server. Set the number of concurrent scripts to run to 1 and maximum number of scripts to run to 2000. Accept the default values for other parameters.

 

Back to Topic

To add variables to the Load Generator script, perform the following steps:

1.

Click the Variables tab.

 

2.

Right-click the Script folder and select Add Variable.

 

3.

Enter var_customerId in the Variable Name field. Verify that the var_customerId variable is set to Integer Range with a minimum value of 1 and maximum value of 2,000 with Sequential access type.

 

4.

Repeat the steps to create two more variables:

Variable Name Content Type Minimum Maximum Access Type
var_reasonCode Integer Range 17 20 Random
var_length Integer Range 75 267 Sequential

 

5.

Repeat the above steps to create another variable var_agent of String Array type with Random access type.

 

6.

Right-click anywhere in the String section and select Add Item.

 

7.

In the enabled row, enter John.

 

8.

Step-off the record and repeat the above steps to add three more random values: Peter, Mary, and Sara.

 

Back to Topic

To add integration point calls to the script, perform the following steps:

1.

Click the Edit Script tab.

 

2.

Right-click anywhere on the left pane and select Add Action.

 

3.

Verify that the action type is set to Message. Enter CallBegin in the Integration Point field.

 

4.

Right-click anywhere in the Input Fields section and select Add Item.

 

5.

Enter customerId in the Name field. Select var_customerId from the Variable field drop-down list. Check the Session Key flag for this variable.

 

6.

Create the ServiceComplete action with the following parameters and map them with corresponding variables.

Name Variables
customerId var_customerId
reasonCode var_reasonCode
agent var_agent
length var_length

Check the Session Key flag for customerId.

 

7.

Create the CallEnd action with the following parameters.

Name Variable
customerId var_customerId

Check the Session Key flag for customerId.

 

8.

Select File > Save to save the script.

 

9.

In the Save window, navigate to Documents and Settings\User \Oracle RTD Studio\Tutorial\etc where Documents and Settings\User is the path to your current windows user, for example, C:\Documents and Settings\JSMITH\Oracle RTD Studio.

 

10.

Enter Tutorial for the filename.

 

11.

Click Save As to save the script.

 

Back to Topic

To run the script and view simulated call reason analysis in Decision Center, perform the following steps:

1.

Click the Run tab.

 

2.

Select Run > Run. This invokes the Load Generator and displays the load status in a graph.

 

3.

Monitor the simulation and verify that the load generator completes the simulations successfully with 0 errors.

 

4.

Select File > Exit to close the Load Generator application.

 

5.

Open Decision Center by opening a Web browser and going to the URL http:// server_name:8080/ui .

 

6.

In the Welcome to Decision Center window, enter sdsu for User name and Password and click Login.

 

7.

Click Open Online Service.

 

8.

In the Select Inline Service window, select the Tutorial Inline Service and click OK.

 

9.

In the Inline Service Navigator, expand the tree to Tutorial (Development) > Decision Process > Call Reason.

 

10.

Expand Call Reason and click the Make Payment choice.

 

11.

Click the Analysis tab.

 

12.

Verify that the Best-fit tab is selected under the Analysis tab. Your results should look similar to the screenshot:

 

13.

In Decision Studio, expand the tree to Tutorial > Service Metadata > Models > Reason Analysis.

 

14.

Double-click the Reason Analysis model to open this window in the editor area.

 

15.

Click the Attributes tab.

 

16.

Click the Select under the Excluded Attributes section.

 

17.

In the Select excluded attribute window, expand the tree to select Session > Call > reason code.

 

18.

Click OK.

 

19.

Select File > Save All.

 

20.

Select Project > Deploy and click Deploy to redeploy the Tutorial project into the RTD server.

 

21.

Verify that you see a "Tutorial deployed successfully" message.

 

22.

Start JConsole by running JDK_HOME/bin/jconsole.exe . (For example, C:\Program Files\Java\jdk1.5.0_08\bin\jconsole.exe).

 

23.

In the JConsole: Connect to Agent screen, click the Remote tab.

 

24.

Enter the host, port, and login credentials for JConsole and click Connect.

 

25.

Click the MBeans tab.

 

26.

Expand OracleRTD > InlineServiceManager > Tutorial > Development > Loadable.

 

27.

Click the Operations tab.

 

28.

Click deleteAllOperationalData( ). This may take a few seconds to complete.

 

29.

Verify that this operation completed successfully with no return values or error message.

Click OK. This deletes all learning and statistics that this service gathered by simulated calls during the Load Generator.

 

30.

Select Connection > Exit.

 

Back to Topic

Back to Topic List

Generating Offers Based on Performance Goals

The goal of this topic is to generate a cross sell offer recommendation based on a performance goal to reduce cost. You have built and tested a basic Inline Service project containing a model to perform call analysis reasons. You are going to enhance the Inline Service project to provide cross-selling advice based on scoring rules to achieve the performance goal of reducing the cost to sell offers.

To generate offers based on performance goals, perform the following steps:

 Create a choice group

Create choices

 Create performance goal
 Create scoring rule
 Assign scores
 Create decisions
 Create an Advisor
 Test

To create a choice group, perform the following steps:

1.

If necessary, start Decision Studio.

 

2.

Expand the tree to Tutorial > Service Metadata > Choices.

 

3.

Right-click Choices and select New Choice Group.

 

4.

Enter Cross Selling Offer in the Display Label field.

 

5.

Click OK. This opens the Cross Selling Offer window in the editor area.

 

6.

Add the following three choice attributes. If necessary, refer to prior topics for steps to add choice attributes.

Attribute Data Type Send to client Overridable
Agent Script String <checked> <checked>
Offer Description String <checked> <checked>
URL String <checked> <checked>

The Add Choice Attribute dialog box is shown only for the Agent Script attribute.

Verify that your work matches the screenshot:

 

7.

Select File > Save All.

 

Back to Topic

To create choices, perform the following steps:

1.

Select the Cross Selling Offer choice group under the Choices folder.

 

2.

Right-click the Cross Selling Offer choice group and then select New Choice.

 

3.

Enter Brokerage Account in the Display Label field.

 

4.

Click OK. This opens the Brokerage Account window in the editor area.

 

5.

Verify that you see all three attributes in the Attributes Values tab.

 

6.

Enter Would you like to try our new brokerage account? in the Attribute Value column for the Agent Script attribute.

 

7.

Enter Brokerage Account offer in the Attribute Value column for the Offer Description attribute.

 

8.

Enter http://www.offer.com/offer1.html in the Attribute Value column for the URL attribute.

 

9.

Repeat the above steps to add four more choices under the Cross Selling Offer choice group:

Choice Agent Script Offer Description URL
Credit Card “Would you like to try our new credit card?” Credit Card offer http://www.offer.com/offer2.html
Life Insurance “Would you like to try our new life insurance?” Life Insurance offer http://www.offer.com/offer3.html
Roth IRA “Would you like to try our new Roth IRA?” Roth IRA offer http://www.offer.com/offer4.html
Savings Account “Would you like to try our new savings account?” Savings Account offer http://www.offer.com/offer5.html

 

10.

Select File > Save All.

 

Back to Topic

To create a performance goal, perform the following steps:

1.

Double-click the Performance Goals group.

 

2.

Click Add.

 

3.

Enter Cost in the Display Label field.

 

4.

Click OK.

 

5.

Set Optimization to Minimize using the drop-down list for this performance goal and accept the default property values.

 

6.

Double-click the Cross Selling Offer choice group.

 

7.

Click the Scores tab.

 

8.

Click Select Metrics.

 

9.

Select the Cost performance metric in the list and click OK.

 

10.

Verify that the Cost performance metric is added to the Cross Selling Offer choice group.

 

Back to Topic

To create a scoring rule, perform the following steps:

1.

Select the Scoring Rules folder.

 

2.

Right-click the Scoring Rules folder and select New Scoring Rule.

 

3.

Enter Credit Card Score in the Display Label field.

 

4.

Click OK.

 

5.

Right-click anywhere in the Value section and select Add Condition.

 

6.

Click the first far left box under the Condition column.

 

7.

Click the Ellipsis button in the far left box.

 

8.

In the Edit value dialog box, check the Attribute option.

 

9.

Navigate to session attributes > customer and then select Age.

 

10.

Click OK.

 

11.

Click the conditional clause and change it to <= using the drop-down list.

 

12.

Enter 40 on the right side of the conditional clause.

 

13.

Enter 130.0 in the Value field next to the conditional clause.

 

14.

Enter 147.0 in the Value field for the Otherwise clause.

 

15. Select File > Save All.

Back to Topic

To assign scores to the cost performance metric of each offer, perform the following steps:

1.

For the Cross Selling Offer choice group, double-click the Brokerage Account choice.

 

2.

Click the Scores tab.

 

3.

Enter 150 as the score for the Cost performance metric.

 

4.

Repeat the above steps to assign scores for each offer as shown below.

Offer Score
Life Insurance 140
Roth IRA 145
Savings Account 135

 

5.

Double-click the Credit Card choice.

 

6.

Click the Scores tab.

 

7.

Click the Score value for the Cost performance metric.

 

8.

Click the ellipsis command button.

 

9.

Check the Function or rule call option and select Credit Card Score from the drop-down list.

 

10.

Click OK.

 

11. Select File > Save All.

Back to Topic

To create new decisions, perform the following steps:

1.

Right-click the Decisions folder under Tutorial > Service Metadata and then select New Decision.

 

2.

Enter Select Offer in the Display Label field.

 

3.

Click OK.

 

4.

Click Select.

 

5.

Select Cross Selling Offer from the list.

 

6.

Click OK.

 

7.

Repeat the steps to create another decision named Random Choice and associate it also with Cross Selling Offer choice group.

 

8.

Check the Select at random check box for the Random Choice decision.

 

9. Save your work.

Back to Topic

To create an Advisor, perform the following steps:

1.

In the Inline Service Explorer, expand the tree to Tutorial > Service Metadata > Integration Points > Advisors.

 

2.

Right-click the Advisors folder and then select New Advisor.

 

3.

Enter Get Cross Sell Offer in the Display Label field.

 

4.

Click OK.

 

5.

In the Request tab, click Select.

 

6.

Select customer / customerId as the session key for this Advisor.

 

7.

Click OK.

 

8.

Set the External System to CRM and Order to 3.

 

9.

Click the Response tab.

 

10.

For Decision, set Decision to Use to Select Offer using the drop-down list.

 

11.

For Control Group Decision, set Decision to Use to Random Choice using the drop-down list.

 

12.

Click Select under the Default Choices section.

 

13.

Select Cross Selling Offer > Life Insurance offer as the default choice.

 

14.

Click OK.

 

15.

Click the Asynchronous Logic tab.

 

16.

Copy the contents from the Lab 9-A section of the Tutorial Code.doc document and paste it in the Asynchronous Logic field.

 

Back to Topic

To test the Advisor logic, perform the following steps:

1.

Select File > Save All.

 

2.

Select Project > Deploy to redeploy the Tutorial project into the RTD server.

Verify that there are no errors in deployment or compilation.

 

3.

In the Test view, select the Service Complete Informant as the integration point from the drop-down list.

 

4.

Enter the following values:

customerId 7
agent John
length 21
reasonCode 19

 

5.

Click the Execute Request button.

 

6.

In the Test view, select the Get Cross Sell Offer Advisor as the integration point from the drop-down list and execute the request with the same customerId value.

 

7.

Click the Response tab.

 

8.

Verify that the selected offer and its attributes are returned and displayed in the Response tab.

 

9.

Verify the log results by clicking the Log tab. The customer's age and presented offer are returned.

 

10.

Start Decision Center and login as sdsu with password sdsu.

 

11.

Click Open Inline Service.

 

12.

In the Select Inline Service window, select the Tutorial Inline Service project.

 

13.

Click OK.

 

14.

Click the Integration Map subtab under the Definition tab. This may take a moment to display.

 

15.

Verify that there are four integration points consisting of three Informants and one Advisor.

Examine the order of these four integration points, which reflects the business process where the agent finishes processing an incoming call in a normal way (marked by Service Complete Informant informs the RTD server) and then wraps up the call presenting a cross-sell offer (using Get Cross Sell Offer Advisor generated by the Inline Service) to the customer.

 

Back to Topic

Back to Topic List

Configuring Inline Service to Learn on Offer Acceptances

The goal of this topic is to configure the Inline Service to learn on offer acceptances. You have built and tested the Inline Service to generate and present a cross-sell offer to a customer based on scoring rules and performance goals. You are now configuring the Inline Service to close the feedback loop by adding a new model that the Inline Service can use to learn on offer acceptances. 

To configure the Inline Service to learn on offer acceptances, perform the following steps:

 Add a model to learn on offer acceptances

Populate the learning model

To add a model to learn on offer acceptances, perform the following steps:

 Create offer events

Create a choice event model

 Configure the session entity to remember the extended offer
 Update the Get Cross Sell Offer Advisor to record the presented offer
 Create an Informant to track the response
 Test the Informant that tracks the response

To create offer events, perform the following steps:

1.

In Decision Studio, double-click the Cross Selling Offer choice group under Tutorial > Service Metadata > Choices.

 

2.

Click the Choice Events tab.

 

3.

Click Add.

 

4.

Enter Presented in the Display Label field.

 

5.

Click OK.

 

6.

Repeat the above step to add another choice event and name it Accepted.

 

7.

Set the Statistics Collector for both events to Choice Event Statistics Collector (using the drop-down list). Accept the default values for Event History (Days) for both choice events, which are set to Session Duration.

 

Back to Topic

To create a choice event model, perform the following steps:

1.

Right-click the Models folder and then New Choice Event Model.

 

2.

Enter Offer Acceptance Predictor in the Display Label field.

 

3.

Click OK.

 

4.

In the Offer Acceptance Predictor window, uncheck the Default time window option.

 

5.

Set the following choice event model properties:

Time Window Week
Choice Group Cross Selling Offer
Base Event Presented
Positive Outcome Events Accepted

 

6.

Click the Learn Location tab.

 

7.

Verify that the learn location is set to On session close.

 

8.

Save your work.

 

Back to Topic

To configure the session entity to remember the extended offer, perform the following steps:

1.

Double-click the Session entity under Tutorial > Service Metadata > Entities.

 

2.

In the Attributes section, click Add Attribute and create a new attribute using the following properties.

Display Label Offer Extended
Data Type String
Show in Decision Center Unchecked
Use for analysis Unchecked

 

3.

Click OK.

 

Back to Topic

To update the Get Cross Sell Offer Advisor to record the presented offer , perform the following steps:

1.

Double-click the Get Cross Sell Offer Advisor located under Tutorial > Service Metadata > Integration Points > Advisors.

 

2.

Click the Asynchronous Logic tab.

 

3.

Replace the contents in the Asynchronous Logic tab with the contents from the Lab 10-A section in the Tutorial Code.doc. At this point, the session knows which offer has been presented to the customer.

 

Back to Topic

To create an Informant to track the response , perform the following steps:

1.

In the Inline Service Explorer, expand the tree to Tutorial > Service Metadata > Integration Points > Informants.

 

2.

Right-click Informants and then New Informant.

 

3.

Enter Offer Feedback in the Display Label field.

 

4.

Click OK.

 

5.

Set the External System to CRM and Order to 4.

 

6.

Click the Select button and set the Session Key to customerId.

 

7.

In the Request Data section, click Add.

 

8.

Enter Positive in the Display Label field and set the data type to String.

 

9.

Click OK.

 

10.

Click the Logic tab.

 

11.

Copy the code from the Lab 10-B section in the Tutorial Code.doc document and paste it into the Logic tab. This logic records the Accepted event when the value of the Informant parameter is "yes."

 

Back to Topic

To test the Informant that tracks the response, perform the following steps:

1.

Select File > Save All.

 

2.

Select Project > Deploy to redeploy the Tutorial project into the RTD server. Verify that there are no errors in deployment or compilation.

 

3.

In the Test view, select Get Cross Sell Offer Advisor as the integration point from the drop-down list.

 

4.

Enter the value 10 for the customerId.

 

5.

Click the Execute Request button.

 

6.

Click the Log tab.

 

7.

Verify that the Get Cross Sell Offer Advisor retrieves and presents the credit card offer.

 

8.

Select Offer Feedback Informant in the Test view.

 

9.

Enter yes for the Positive input parameter.

 

10.

Click the Execute Request button.

 

11.

Verify that the Offer Feedback Informant records the acceptance of credit card offer. This verifies that your closed-loop decision process works. The Get Cross Sell Offer Advisor extends and presents an offer to the customer. The Offer Feedback Informant records when and whether the offer was accepted. The Offer Acceptance Predictor model learns based on presented and accepted events.

 

Back to Topic

To populate the learning model, perform the following steps:

 Create the Load Generator script

Remove the existing simulated learning and statistics

 Run the script and view simulated call reason analysis results in Decision Center

To create the Load Generator script, perform the following steps:

1.

Start the Load Generator.

 

2.

Click Open an existing Load Generator script option from the list.

 

3.

In the Open window, navigate to Documents and Settings\User \Oracle RTD Studio\Tutorial\etc.

 

4.

Select Tutorial.xml. This is the same script you built in a previous exercise.

 

5.

Click Open.

 

6.

Click the Variables tab.

 

7.

Right-click the Script folder and select Add Variable.

 

8.

Enter var_positive in the Variable Name field.

 

9.

Select Weighted String Array in the Contents field.

 

10.

Right-click anywhere on the String section and then select Add Item.

 

11.

Enter 30 for the Weight and yes for the String. The Load Generator returns this string value 30% of the time.

 

12.

Repeat the above steps to add another element with a value of 70 for the Weight and no for the String. The Load Generator returns this string value 70% of the time.

 

13.

Click the Edit Script tab.

 

14.

Right-click anywhere on the left pane and select Add Action.

 

15.

Verify that the action type is set to Message.

 

16.

Enter GetCrossSellOffer in the Integration Point field.

 

17.

Right-click anywhere in the Input Fields section and select Add Item.

 

18.

Enter customerId in the Name field.

 

19.

Click the Variable field.

 

20.

Select var_customerId from the drop-down list.

 

21.

Check the Session Key flag for this input field.

 

22.

Right-click the GetCrossSellOffer action and select Move Up to move up one position.

 

23.

Repeat the steps to create another integration point, OfferFeedback.

 

24.

Add two input fields customerId and positive.

 

25.

Set the customerId input field to the var_customerId variable and make this input field the session key for this action.

 

26.

Set the positive input field to var_positive variable.

 

27.

Move up the OfferFeedback action so that it is right after the GetCrossSellOffer action but before the CallEnd action.

 

28.

Select File > Save to save the script.

 

Back to Topic

To remove the existing simulated learning and statistics, perform the following steps:

1.

Start JConsole by running JDK_HOME/bin/jconsole.exe .

 

2.

In the JConsole: Connect to Agent screen, click the Remote tab.

 

3.

Enter the host, port, and login credentials for JConsole and click Connect.

 

4.

Click the MBeans tab.

 

5.

Expand OracleRTD > InlineServiceManager > Tutorial > Development > Loadable.

 

6.

Click the Operations tab.

 

7.

Click deleteAllOperationalData( ). This may take a few seconds to complete.

 

8.

Verify that this operation completed successfully with no return values or error message.

Click OK. This deletes all learning and statistics that this service gathered by simulated calls from the Load Generator.

 

9.

Close JConsole.

 

Back to Topic

To run the script and view simulated call reason analysis results in Decision Center, perform the following steps:

1.

In the Load Generator, click the Run tab.

 

2.

Select Run > Run. This invokes the Load Generator and displays the load status in a graph.

 

3.

Monitor the simulation. This script is going to simulate the scenario where offers are extended 2000 times and they are accepted 30% of the time. Verify that the script completes with no errors.

 

4.

Start Decision Center and log in as sdsu with password sdsu.

 

5.

Click Open Inline Service.

 

6.

In the Select Inline Service window, select the Tutorial Inline Service project.

 

7.

Click OK.

 

8.

Click the Cross Selling Offer choice group.

 

9.

Click the Performance tab.

 

10.

Verify that the Counts subtab is selected. Your results should look similar to the screenshot.

Notice that two of the five offers, Credit Card and Savings Account, are extended. The offer decision so far is based on the performance goal of minimizing the cost of the cross-sell offer. The score value assigned to these offers is such that depending on the customer’s age the lowest cost offer was either Credit Card or Savings Account. Either one of the offers is accepted approximately 30% of the time. The 30% offer acceptance percentage is the predicted probability because you set the Informant OfferFeedback to receive the Accepted event 30% of the time using the var_positive variable in the Load Generator utility. Since this is a randomized event, your results may not be exactly 30% but should be close to this percentage.

 

Back to Topic

Back to Topic List

Using the Model to Influence Offer Generation

To use the model to influence offer generation, perform the following steps:

 Expand the performance goal

Populate the model to influence offer generation

To expand the performance goal, perform the following steps:

 Create scoring rule for the Expected Revenue performance goal

Create a performance goal to maximize the expected revenue from the offer

 Update the decision process to include Expected Revenue performance metric
 Check and view values for the likelihood of acceptance
 Test the Advisor logic

To create a scoring rule for the Expected Revenue performance goal, perform the following steps:

1.

In Decision Studio, double-click the Cross Selling Offer choice group.

 

2.

In the Choice Attributes tab, add a new attribute Revenue. Set the data type to Integer for this attribute and check the Send to client option.

 

3.

Click OK.

 

4.

Expand the Cross Selling Offer choice group.

 

5.

Double-click the Brokerage Account choice.

 

6.

In the Attribute Values tab, set the Revenue attribute value to 300.

 

7.

Repeat the above steps to set the Revenue attribute value for other offers as listed below.

Credit Card 205
Life Insurance 185
Roth IRA 190
Savings Account 175

Only the Savings Account offer is shown in the screenshot:

 

Back to Topic

To create a performance goal to maximize the expected revenue from the offer, perform the following steps:

1.

In the Inline Service Explorer, expand the tree to Tutorial > Service Metadata.

 

2.

Double-click the Performance Goals folder.

 

3.

Click Add.

 

4.

Enter Expected Revenue in the Display Label field.

 

5.

Click OK. Verify that Optimization for the Expected Revenue performance goal is set to Maximize. Accept the default property values for this performance goal.

 

6.

Double-click the Cross Selling Offer choice group.

 

7.

Click the Scores tab.

 

8.

Click Select Metrics and add the Expected Revenue performance metric to the scores list. The Cross Selling Offer choice group now contains both Cost and Expected Revenue performance metrics.

 

9.

Click File > Save All.

 

10.

Click the Score column value for the Expected Revenue performance metric.

 

11.

Click the Ellipsis button on the Score column.

 

12.

Select Function or rule call value source and select Multiply from the Function to Call drop-down list.

 

13.

Click the Value column for Parameter a, and then the Ellipsis button.

 

14.

In the Edit Value window, select Attribute or variable value source.

 

15.

Select Revenue under the Choice folder.

 

16.

Click OK.

 

17.

Repeat the steps for Parameter b by selecting the Model prediction value source and setting the Accepted event from the Offer Acceptance Predictor model.

 

18.

Click OK to close the Edit Value window of Parameter B.

 

19.

Click OK to close the Edit Value window of Expected Revenue performance metric.

Examine the score logic for the Expected Revenue performance metric. The score for the Expected Revenue performance metric is a product of Revenue and Likelihood of acceptance. The Likelihood of acceptance is a number from 0 to 1, where 1 means 100% likely to accept, generated by the model and specific to each offer. If offer A and B both have the same base revenue, then the revenue score will be higher for the offer that has a higher likelihood of acceptance value.

 

Back to Topic

To update the decision process to include theExpected Revenue performance metric, perform the following steps:

1.

In the Inline Service Explorer, expand the tree to Tutorial > Service Metadata > Decisions.

 

2.

Double-click the Select Offer decision.

 

3.

Click Goals under Target Segments section.

 

4.

Select the Expected Revenue performance metric in the Select Performance goals window.

 

5.

Click OK.

Upon adding the Expected Revenue performance metric, the weight is automatically distributed across available performance goals (50% each by default).

 

Back to Topic

To check and view values for the likelihood of acceptance, perform the following steps:

1.

Double-click the Cross Selling Offer choice group.

 

2.

Click the Choice Attributes tab.

 

3.

Add a new attribute, Likelihood of Acceptance, to this choice group with its data type set to Double and with the Send to client option selected.

 

4.

Set the Likelihood of Acceptance attribute value by selecting Model Prediction as the value source and choosing the Accepted event from the Offer Acceptance Predictor model.

 

5.

Double-click the Get Cross Sell Offer Advisor.

 

6.

Click the Asynchronous Logic tab.

 

7.

Replace the code in the Asynchronous Logic tab by copying and pasting the contents from the Lab 11-A section of the Tutorial Code.doc document. This modified code prints the likelihood of acceptance value to the server log.

 

Back to Topic

To test the Advisor logic , perform the following steps:

1.

Select File > Save All.

 

2.

Select Project > Deploy to redeploy the Tutorial project onto the RTD server. Verify that there are no errors in deployment or compilation.

 

3.

In the Test view, select the Get Cross Sell Offer Informant as the integration point (from the drop-down list) and enter 10 for the customerId.

 

4.

Click the Execute Request button.

 

5.

Click the Response tab.

Verify that the selected offer and its attributes are returned and displayed in the Response tab. It is okay if you see a NaN (Not a Number) for the likelihoodofAcceptance attribute. This means the model does not have enough data to compute the likelihood value for this offer. The number of iterations necessary to reach model convergence to predict likelihood of acceptance depends on the application and quality of the data.

 

6. Introduce offer acceptance bias. At this point in time, all offers have a 30% chance of being accepted and there will not be any strong correlations between customer profile and offer acceptance. You are going to introduce offer acceptance bias to explore how the model results will change because of this bias. The artificial bias will record offer acceptance for customers who have two or more children with Life Insurance offer.

Double-click the Offer Feedback Informant.

 

7.

Click the Logic tab.

 

8.

Replace the code by copying the contents from the Lab 11-B section in the Tutorial Code.doc document and pasting it into the Asynchronous Logic tab. This modified code records the feedback that an offer has been extended and accepted by customers who have two or more children. Please note the comment in the code:

If data source is Oracle, change the line

int numOfChildren = session().getCustomer(). getNumberOfChildren();

to

int numOfChildren = session().getCustomer(). getNumberofchildren();

 

9.

Select File > Save All.

 

10.

Select Project > Deploy to redeploy the Tutorial project onto the RTD server. Verify that there are no errors in deployment or compilation.

 

Back to Topic

To populate the model to influence offer generation, perform the following steps:

 Remove the existing simulated learning and statistics

Run the script using the Load Generator utility

 View the results of artificial bias into the model in Decision Center

To remove the existing simulated learning and statistics , perform the following steps:

1.

Start JConsole by running JDK_HOME/bin/jconsole.exe .

 

2.

In the JConsole: Connect to Agent screen, click the Remote tab.

 

3.

Enter the host, port, and login credentials for JConsole and click Connect.

 

4.

Click the MBeans tab.

 

5.

Expand OracleRTD > InlineServiceManager > Tutorial > Development > Loadable.

 

6.

Click the Operations tab.

 

7.

Click deleteAllOperationalData( ). This may take a few seconds to complete.

 

8.

Verify that this operation completed successfully with no return values or error message.

Click OK. This deletes all learning and statistics that this service gathered by simulated calls from the Load Generator.

Back to Topic

To run the script using the Load Generator utility , perform the following steps:

1.

If necessary, start the Load Generator utility.

 

2.

Click Open an existing Load Generator script option from the list.

 

3.

In the Open window, navigate to \Oracle RTD Studio\Tutorial\etc.

 

4.

Select Tutorial.xml.

 

5.

Click OK.

 

6.

If necessary, clear the graphs by selecting View > Clear Graphs.

 

7.

Click the Run tab.

 

8.

Select Run > Run. This invokes the load generator and displays the load status in a graph. 

 

9.

After about 200 total finished scripts, click the pause button to temporarily stop sending requests to the server. View the server’s output in the server.log file. You will see that the printed Likelihood Of Acceptance values are NaN for all sessions. This is an indication that the model has not yet learned enough data to be able to compute the likelihood of acceptance. Note that offers are still being presented despite the lack of likelihood values. Offers are being selected using built-in scores comparison logic.

 

10.

Un-pause the Load Generator script and let it finish running for 2000 total finished scripts. In the server output, you should now see actual values for Likelihood Of Acceptance, varying around 0.3 for all offers except Life Insurance, which has higher values because of the bias introduced.

 

11.

It is important to note that the model-predicted Likelihood Of Acceptance values for a given offer will differ for different customer profiles. For example, suppose we have two customers John and Tom, who only differ in the number of children they have. If we printed the Likelihood Of Acceptance values for the Life Insurance offer for these two customers (at a snapshot in time), we will see a higher value for Tom. This is because Tom has three children, and is therefore more likely to accept the Life Insurance offer, if it is presented to him.

Customer
# of children
Likelihood Of Acceptance for offer Life Insurance
John Doe
0
0.32
Tom Smith
3
0.89

Since we determine which offer to present to the customer based on the combination of Cost and Maximize Revenue scores, and because Maximize Revenue depends on the model’s predicted Likelihood Of Acceptance value for each offer, the Life Insurance offer will have a high Maximize Revenue value for customers with two or more children, and therefore for such customers, Life Insurance will be presented (and then accepted) far more frequently than other offers!

 

Back to Topic

To review the results of artificial bias into the model in Decision Center , perform the following steps:

1.

Start Decision Center and log in as sdsu with password sdsu.

 

2.

Click Open Inline Service.

 

3.

In the Select Inline Service window, select the Tutorial Inline Service.

 

4.

Click OK.

 

5.

Double-click the Cross Selling Offer choice group.

 

6.

Click the Performance tab.

 

7.

Verify that the Counts subtab is selected.

Notice that all five offers are extended. The average percentage of times these cross-sell offers are accepted is 30% for all offers except Life Insurance. The offer acceptance percentage for the Life Insurance offer is higher than the rest of the offers. The inclusion of an artificial acceptance bias is the reason for this phenomenon. Customers with two or more children are automatically set to offer the Life Insurance offer.

 

8.

Double-click the Life Insurance choice.

 

9.

Click the Analysis tab.

 

10.

Click the Best-fit tab.

Notice that the value for the Model Quality is over 50. Notice also that the the highest correlating attribute for the Life Insurance offer is customer NUMBEROFCHILDREN, whose value is 2. This is fully expected because of the artificial bias to present this Life Insurance offer to customers with two or more children.

 

Back to Topic

Back to Topic List

Summary

In this lesson, you learned how to:

 Create a New Inline Service
 Create Models for Call Analysis
 Generate Offers Based on Performance Goals
 Configure Inline Service to Learn on Offer Acceptances
 Use a Model to Influence Offer Generation

 

Back to Topic List

Related Information

To learn more about Oracle RTD you can refer to:

 Additional OBEs on the OTN Web site.
 Oracle RTD on Oracle.com

Back to Topic List

 Place the cursor over this icon to hide all screenshots.

 

 

 

Left Curve
Popular Downloads
Right Curve
Untitled Document