We're making the ticket purchasing experience quicker, more intuitive, and more accessible. Using voice analysis we are able to provide travelers with exactly what they want and in a matter of seconds, rather than wasting their time with clunky interfaces that create stress and long queues.
Train of Thought is built by integrating advanced speech recognition API's, IBM's Watson Artificial Intelligence services including the Natural Language classifier and the SilverRail API to pull up to date information about train routes and schedules in the United Kingdom. As a voiced based application, we focussed on an ultra-high accuracy of speech-to-text transcription using Google's new web speech API. The accuracy is above 90% for most accents and use cases, and we achieved 100% accuracy for most tests and common queries. IBM's Natural Language classifier service, a part of the Bluemix basket of Watson service, then run the transcribed speech through a classifier that we trained to recognize common customer queries at kiosks such as route finding, ticket purchasing and ticket printing based on a pre-booked purchase and reference number. The confidence index for most results returns by the classifier is over 98% due to the extraordinarily well trained, machine learning service that we built and integrated. Once queries are classified, the Dialog AI service from Watson is used to provide a fast and accessible interaction with the customer to buy tickets, see route information or print pre-booked tickets. For ticket purchasing and routing, the SilverRail API is called to get the up to date routes between two locations which are extracted from the transcribed text.
A seamless, fast and accessible user experience. Buy a ticket in less than 10 seconds.
Built With
- google-voice-api
- ibm-watson
- silverrail-api
Log in or sign up for Devpost to join the conversation.