AI Case Study
Trainline automates disruption information and alerts using natural language processing
Trainline has introduced a new system to automate disruption information updates and alerts. Using natural language processing and machine learning the system analyses rail operators' Twitter accounts feeds to gather the relevant data on disruptions. A notification system classifies the messages per importance and then the stations that are affected by a disruption are determined through contextual scoring. Individual customers whose journeys might be affected get updates through the Trainline voice app, possibly before the national rail data update its feeds.
Consumer Goods And Services
Travel And Leisure
"Trainline is using artificial intelligence (AI) to automate disruption alerts in its voice app for Google Assistant.
The Trainline team's model trawls rail operators' Twitter accounts for relevant data. The information is then available to passengers when they ask questions like 'How's my commute doing?' or 'Is this train on-time?'
The system uses natural language processing and machine learning to analyse the data on rail disruptions collected from Twitter feeds. The notification system automatically classifies the importance of a message, while a second layer of contextual scoring determines which stations are affected by a disruption.
The AI automatically matches this information to individual user journeys and uses it to inform customers using the voice app about disruptions. Trainline says that alerts will often be pushed "before this data is available through the national rail data feeds".
Customers can also view the history of the disruption, so they can see its scale, when it started, how it has unfolded and what is being done to fix it.
The voice disruption notification feature is available in the Trainline voice app for Google Assistant now"
"The system uses natural language processing and machine learning to analyse the data on rail disruptions collected from Twitter feeds. The notification system automatically classifies the importance of a message, while a second layer of contextual scoring determines which stations are affected by a disruption."
Technical And Product Support
Rail operators' Twitter feeds