We were tasked with creating a hybrid mobile application that would offer standard Customer Relation Management functionalities, empowered with machine learning solutions, and innovative functionalities to guarantee integrity of contact data.
A goal was to support natural language processing, voice chat, and interpret commands with the use of machine learning. Integrity of contact data was planned to be guaranteed with the use of a blockchain technology, which would allow to synchronize dispersed knowledge in community of sellers.
Functionalities were fully synchronized with SalesForce, Google Calendar and Google Contact, with plans to introduce additional synchronizations. The application was based on a pre-existing code, which we had to refactor and replan.
We started the project with a technological audit which pointed out what needed to be improved in the application’s architecture. We adviced to rewrite front-end to React. We used an architecture of loosely coupled services to improve future-proofness of the back-end.
Once the application had been refactored, we started implementation of new functionalities.
The ability to transform speech to text on mobile devices was guaranteed by a custom cordova plugin, built as a wrapper for Microsoft iOS Bing SDK. We implemented machine-learning functionalities with the use of Microsoft Cognitive Services - mainly LUIS. This included eg. an automatic extraction of new entities and activities from emails that were processed by the system. We implemented an intelligent SalesBot with the use of Microsoft Bot Framework.
We implemented a role management system, and with the use of GraphQL we started creating a single point of truth for contact data. Afterwards we added a synchronization of data with Google Contacts and Google Calendar, as well as synchronization with Salesforce.