Smart Advice. Agile. Personalized. Transparent.

GA Release for Intelligent Advisor 20C (12.2.20)

Brad Tuckett
Release Manager

Intelligent Advisor Release 20C (12.2.20) is now available for public cloud and private cloud (on-premise) customers.

Develop decision services in a browser (Monthly update 2)

Using Intelligent Advisor Hub you can now create and deploy decision-making services in a web browser. This alleviates the need to install and manage a desktop development tool (Policy Modeling). Browser-based rule authoring centralizes the custom application logic and enables the decision-making rules to be easily shared with stakeholders.

The inputs and outputs that the decision service expects are defined and managed separately from the decision logic in a decision service project. That is, only users with the Manager role can modify the input and output contract, and activate published versions. Users with the Author role can make any other changes, and publish and test those changes. This separation enables business users to quickly make changes knowing that they can’t accidentally break integrations that are expecting to pass in and receive back particular inputs and outputs.

The rules are written in natural language, and when published, a self-documenting web service is automatically created using Open API format. The decision service can be consumed from an Oracle Visual Builder application or any other application framework.

The project details view for a decision service project in Intelligent Advisor Hub

Ability to create additional deployment statistics charts (Monthly update 1)

Hub Administrators can now create more than three statistics charts, both for the whole Intelligent Advisor Hub site, and for specific deployments.

Chat API enabled for OPA for Workers SKU users (Monthly update 1)

The Intelligent Advisor Chat API can now be used on Intelligent Advisor sites that are licensed via the OPA for Workers SKU. This enables Intelligent Advisor interviews to be delivered to employees as self-service chatbot experiences.

Auto fill input purpose

Input fields can now be tagged with an auto fill input purpose so that browsers can use stored data, such as the person's name and email address, to pre-populate information in an interview. The stored data in the browser may come from previously entered values or from pre-configured values (for example, in Google Chrome, the data stored in the user's Auto-fill settings). This automated assistance reduces typing and enables Intelligent Advisor interviews to be Web Content Accessibility Guidelines (WCAG) 2.1 compliant.

Test case execution report

When all the test cases are run in a project, two reports are now automatically generated:

  • a TestCaseLog CSV file - this records the time taken to run each test case
  • a testPerformance CSV file - this records the time spent testing each individual rule.

Comparing the information in these reports across versions allows early detection of project changes that may have inadvertently caused slowdowns in overall performance. Test coverage can also be improved by filling gaps where reports show rules are not currently being tested.

Rule profiling report

Rule authors can now run batches of data through their rules and get a report that shows where time is spent processing the request. This rule execution profiling report enables authors to decide whether alternate data structures or decisioning approaches might yield better batch processing throughput.

Batch timing enhancements

System integrators can now see in the response from the batch assess service:

  • how long was spent processing each case in a batch request (shown in the @time field), so that they can analyze how data size or other factors affect performance, and
  • how long each batch request spent queued before processing started (shown in the processorQueuedSec field), so that they can understand whether they are pushing too many concurrent batch requests to the server in order to optimize the total load sent.

Branch a repository project

Project authors can now branch a project in Intelligent Advisor Hub to create a separate project version that can be worked on independently. This allows ongoing development to continue in one version of the project while another version is being finalized for release. A link is automatically created in each branched project that takes the user directly back to the version the project was branched from. This enables authors to understand the lineage of projects to ease future change tracking and troubleshooting.

Discard invalid data when interview loads

Interview designers can now specify that they would like any loaded data that is invalid discarded when the interview starts. This enables advice experiences to be more resilient by tolerating mismatches between application metadata (such as a list of valid options) and existing system data.

Component versioning

Intelligent Advisor now has a component version (for example, 42.0.0) that is distinct from the product version (for example, The component version is the version of the interface between Oracle Policy Modeling and Intelligent Advisor Hub. It is independent of the product version, meaning that it could change during a release without the product version changing, and that it might not change even though the product version changes. You can see the component version on the Welcome page of Policy Modeling.

Want to start using the 20C release?

Public cloud: You can upgrade your testing sites at any time via Oracle Cloud Portal Service Cloud Configuration Assistant. All sites will be automatically upgraded according to the published schedule.

Private cloud (on-premises): Licensed customers should download from Oracle® Software Delivery Cloud.  Developers can download and evaluate recent releases from the Oracle web site.

Want further information?

A summary presentation of the key features and benefits in this release is available on the Oracle web site.

Be the first to comment

Comments ( 0 )
Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.