Using the SAS Intelligent Decisioning Connector in SAS Customer Intelligence 360
Introduction
SAS Intelligent Decisioning is a powerful solution to streamline, automate and enhance your decision-making processes. In this post, we will show you how to leverage this capability through SAS Customer Intelligence 360. We will walk you through the steps to set up a SAS Intelligent Decisioning connector, ensuring you can integrate decisioning into your customer analytics solutions.
Understanding the SAS Intelligent Decisioning Connector
SAS Intelligent Decisioning is a software solution that helps organisations automate and manage decisions by combining business rules, custom code, and machine learning models into a single decision flow. It is also possible to have branches to different decision paths depending on the value of certain variables. Each decision flow has its own input variables and output variables associated with it.
Input variables are those required at the start of the flow. Output variables are those available at the end of the flow. These can either be input variables that flow through the process or new variables (such as scores or decisions) that are created during the flow.
Select any image to see a larger version.
Mobile users: To view the images, select the "Full" version at the bottom of the page.
The SAS Intelligent Decisioning Connector enables you to pass real-time behavioural data from CI360 to decision flows in SAS Intelligent Decisioning. The output variables are then passed back to CI360 via the connector and can then be used to enhance targeting and personalization of content. For example, a marketer might use the Accept/Decline variable to show different creatives appropriate for that decision.
The connector references the decision flows that you define in SAS Intelligent Decisioning. Each decision flow is its own endpoint in the connector. For each endpoint, there are one or more trigger events associated with it. For example, this could be a page view or button selection. When this event happens, a request call is sent to SAS Intelligent Decisioning. The request call contains all the input variables required by the decision flow.
Pre-requisites
Several steps must be carried out before the connector can be set up:
- Any data being passed to Intelligent Decisioning must first be registered in the Customer State Profile. This is an up-to-date view of important information (including behaviors, and customer data) that can be accessed to satisfy real-time use cases for Intelligent Decisioning. Standard events that are generated by the system can be registered through the Customer State Management page. This can be found by going to “Administration” then “General Settings” and the “External Access” section. Other events can be registered by going to the Event home page that you want to register. Once here, we click on the ellipses at the top right to bring up the drop-down menu. We then click on “properties” and tick the check box that registers the event.
- You must ensure that you have a client id and secret key for the Viya installation where Intelligent Decisioning resides. Both these tasks need to be done through the administrator set-up for SAS Viya. More information can be found at the SAS Developer website: https://developer.sas.com/rest-apis#authentication-to-sas-viya
- You must first define the decision flows in SAS Intelligent Decisioning that you wish to use. Please ensure you have done this before proceeding.
Step-by-Step Guide to Setting Up the Connector
In this demonstration, we will use a pre-existing decision flow which determines if home insurance can be offered, and if so, what the premiums would be. We also have a pre-existing webform for the web user to fill in details about the property they wish to insure.
To create a SAS Intelligent Decisioning connector we need to go to the main connector page. This can be found in the External Access section under General Settings. We then click on the New Connector icon at the top right.
This brings up the New Connector window. We select the Intelligent Decisioning Connector from the list and click the “Next” button.
This brings up the window for the connector definitions.
We give the connector a name and an optional description. If you need an access point, it must be already configured and available in the drop down. The last field is the support contact field. This should contain email addresses or distribution lists that will be automatically notified if errors occur while the system is processing requests to SAS Intelligent Decisioning, or if the connector invocation fails. Multiple entries can be added by using semi-colons. This field is normally mandatory, but you can skip this part if you use the tick box and opt out of email notifications. Care should be taken when opting out, as it means there won't be any notification of errors. Users will solely be reliant on the Connections Dashboard for alerts. For this reason, the opt out is often only used during the connector build and testing phase.
Once we have completed the definitions, we click on the “Next” button. We then see the window to specify the connector settings.
In this window, we first enter the URL that points to the SAS Micro Analytic Service engine of SAS Intelligent Decisioning. Note that server certificate validation is enabled to enhance the security of the connection. The TLS server certificate is verified to ensure that the connection to the server has direct end-to-end protection for HTTPS protocol. Please ensure that you use a valid SSL certificate. Otherwise, the connector might not function.
We also enter the client ID and secret key that we obtain from the points discussed in the pre-requisites section. Once all this information is complete, we click on “Return Endpoints” to return all available endpoints from this instance of SAS Intelligent Decisioning. Each decision flow will create one endpoint.
Configuring Endpoints
All returned endpoints are unconfigured and disabled by default.
To configure them, we need to add a trigger event and the inputs for the decision. We click on the endpoint name and add this information in the endpoint settings window. We start with the orchestration tab. This specifies the trigger events.
We first ensure that the “Select Events” radio button is activated. Then we click "Add" which brings up the list of all available trigger events. We select one or more events, click “OK” then click “Apply”. In this instance, we will use the completion of the web form as the event trigger.
We then go to the Details tab. The input and output variables for the endpoint (i.e. the decision flow) are automatically listed here. We now need to map the decision inputs to data from CI360.
If we click on the Data Items box, we can see the list of data items from the customer state profile. Note that if we want to use event attributes from the trigger event, we must enter it manually as a JSON path format. This will be of the form: $.properties.“attribute name”. All the decision inputs we need are coming from the web form trigger event, so we have used the JSON path format in all the mappings. You can optionally also enter a default value and a description for each input.
We specify a data label for the output variables. This data label will then be visible in the "Decision Flows and Models" folder that is displayed in the list of data items that you can select for targeting or personalisation. You can optionally provide a description for the output variables too. When you have added all the input, output and trigger events, you have then completed the configuration. You can then click “Apply” to save the configuration.
We then see the endpoint has been configured with the green tick icon. We must also enable the endpoint using the slider button. They are disabled by default. Once enabled, we can also test the endpoint.
We click on the endpoint name to open it and scroll down to the Test Endpoint section at the bottom. We then enter the complete JSON request body with some sample values in the properties section. Note that you cannot view the Test Endpoint section if the endpoint is disabled or unconfigured.
To automatically format the JSON script, click “Format”. This applies the correct spacing and indentation. Once its ready, click on “Run Test”. When complete, you will see the decision response in the Test Results section. We can also see the status code, time elapsed and the size of the returned payload.
The test is successful. I can see there is a status code of 200 and the values returned for the premiums are correct. The endpoint is now ready for use. For example, I could use this to personalise messages to web users that have completed the webform.
Conclusion
Integrating CI360 with SAS Intelligent Decisioning is a great way to leverage powerful decisioning capabilities to enhance marketing campaigns. The built-in SAS Intelligent Decisioning connector available in CI360 delivers this integration with just a few button clicks. By following the steps outlined in this post, you can set up these connectors quickly and efficiently.
There is an upcoming demonstration of this connector in the SAS Customer Intelligence 360 Learning Subscription for Customers. This will be found in the “Using Agents and Connectors with SAS Customer Intelligence 360″ section of the course.
Find more articles from SAS Global Enablement and Learning here.
