Predict Mobile is a system that helps organisations find the best possible price and service combination for their corporate mobile contracts - leading to average savings of 35% which can be very significant saving on major corporate mobile accounts. Ember was instrumentally involved in helping Predict Mobile secure a Scottish Enterprise SMART Award to research and develop the AI elements of this system and to design a technical Architecture & Hosting environment capable of handling the volume of data required to crunch the numbers.
We worked with Data Analysts & AI specialists at the University of St Andrews to develop the predictive elements of the system using the Tensor Flow Framework and we used Selenium Scripting to automate the training of the AI model with high volumes of user simulated data.
We designed an onboarding and brokering system capable of taking both corporate users and business mobile providers through complex user journeys such as uploading billing data and responding to contract tenders by constantly presenting compelling value and progress throughout each workflow.
The system analyses historical billing data and combines this with a mix of client specific and generalised industry future expansion considerations - simultaneously optimising the structure and price of a future mobile contract provision by minimising the potential for both unused service or out of bundle charging.
This information package is automatically converted into an optimised request for pricing that's distributed to registered business mobile contract providers, seeking the best combination of price and service for the client.