Architecture

Overview

eXvisory has a simple cloud-based architecture.

eXvisory core architecture

eXvisory dev is a visual web editor for building deep logic networks capable of troubleshooting complex problems. When you Save from eXvisory dev it generates Java 'engine' code that implements the complex diagnostic reasoning captured by your deep logic network. If you inspect the Java code you'll see it's made up of thousands of interconnected IF...THEN statements. It's easy enough to follow through the code in a debugger but it would be a nightmare to build, maintain and test it manually.

The generated Java 'engine' code is automatically made available via a cloud-hosted REST API, which is used by our web-hosted chatbots and other chatbot frameworks. When you save your deep logic network from eXvisory dev, e.g. at https://dev.exvisory.ai/apps/sample-mobile, it is immediately available as a web chatbot at https://dev.exvisory.ai/apps/sample-mobile.

Natural language

The eXvisory architecture only uses natural language understanding (NLU) to provide automatic translation of our web-hosted chatbots into other languages. It does NOT provide the NLU capabilities other chatbot frameworks like IBM Watson and Google DialogFlow provide, which enable users to ask free-form questions. eXvisory complements NLU frameworks by adding deep logic reasoning, in much the same way that a medical expert must both understand what his/her patients are asking but also have deep medical diagnostic expertise.

The eXvisory REST API can be used to add deep logic skills to NLU chatbot frameworks like IBM Watson and Google DialogFlow. See our article Shallow chatbots are boring: adding deep logic to Google DialogFlow to see how it works and contact us if you want to try it out!

‚Äč