As part of Quest’s 2019 Orchestrator Week, Noel D’Costa, Principal Software Engineer at Oracle, gave a presentation about how to leverage emerging technology like machine learning, robotic process automation, and chatbots in JD Edwards.

Digital Assistants and Chatbots in JD Edwards

The Current State of Chatbots

Until now, chatbots only solved simple business problems like customer service automation, automation of structured processes, and self-service. Today’s digital assistants are more user-oriented and can assist users with various related or unrelated tasks. A digital assistant is a computer program that uses artificial intelligence (AI) and machine learning (ML) to combine independent chatbots into a single conversational interface that assists real users throughout their day.  Their conversations are context-aware, and they learn about the user as they interact.

Oracle Digital Assistant

Oracle Digital Assistant core modules include:

• Skills: Individual chatbots that are reusable, leverage conversational AI and ML, and work with any messenger
• Digital Assistants: A “master bot” that invokes skills, performs smart routing, and disambiguates bot requests if needed
• Channels: Include messenger connectivity, webhook, and native channels (iOS, Android, Web, SMS, Facebook, and WeChat)
• Store: Similar to an App Store where bots from Oracle SaaS are available

Skills include intent resolution, entity extraction, and conversation flows. Each skill usually relates to a business domain like finance, HR, or room booking. Intents implement use cases within skills such as ordering a pizza, canceling a pizza, tracking an order, or showing a menu. User input is interpreted by a trained model where intents are resolved and mapped to a dialog. Entities are variable elements of intent such as pizza size, topping, delivery address, or delivery time. Conversation flows define the chat where the interaction between user and skill results in the completion of a task.

Human agent interaction is one way in which bots are used. Chatbots are often used in the first level of support. They answer frequently asked questions and assist with low tech engagements. They also can hand the user off to a human agent.

Another skill is Embedded Intent Tester. This skill allows batch testing, which is very helpful after model retraining. It displays extracted entities and values and is a natural language understanding tester. It tests the confidence level to which intent is resolved.

Backend integration is an additional skill. There are remote services integration through custom components. It is a component-based approach, which integrated with data and business services and is a “write once and use anywhere.” There are also custom components, which allows for writing direct response messages, save data in variables, and influences conversation flow.

Digital assistants also have the option of adding skills. Published skills are displayed in the skill catalog. Configured skills indicate when a new version is available. Requests are routed to best fit skill model and flow. The user can set thresholds for identifying skills and matching intents.

Channels connect skills and digitals assistants to messenger clients. Native channel support provides configuration for specific messengers. Webhook channel allows messengers with no native channel support to interact with digital assistant or skill.

For a demonstration of digital assistants in JD Edwards, check out the video below.

How to Build Your Own Chatbot in JD Edwards

Follow these steps to build your own chatbot in JD Edwards:

1. Use JD Edwards Orchestrator Studio to develop orchestrations to automate the steps required to perform a business process or action.
2. Use Oracle bot SDK/NodeJS to create the custom component and connect the bot to JD Edwards.
3. Use Oracle Digital Assistant Cloud Service to create the intents and entities and define the flow to invoke the custom component.
4. Use NodeJS chatbot web UI sample JD Edwards EnterpriseOne Page Manager to package and upload the bot UI as an EnterpriseOne page.
5. Use JD Edwards to run it.

To learn more about how to leverage digital assistants and chatbots in JD Edwards, check out the 2019 Orchestrator Week presentation below. To access all of the recordings from Orchestrator Week, check out the Orchestrator Week landing page.