New Transport for London social media tool uses artificial intelligence for chat facility with customers
London’s public transport authority has launched a new chat function that harnesses artificial intelligence (AI) through Facebook Messenger.
Transport for London (TfL) said the TravelBot can respond to customer enquiries through the text message platform, with the possibility of linking direct to a customer service agent.
A spokesperson for the organisation told UKAuthority that it works through the Google+ API service, and provides more flexibility than previous online enquiries by responding to the customer’s choice of words in the message.
TfL said that as people use the service, it will learn and become even more precise.
Existing features include the capacity to provide service updates on bus routes, the Underground, TfL Rail, London Overground, the Docklands Light Railway and London Trams. It can also provide information on bus arrivals, Underground and rail maps, and refer customers to a service agent.
The organisation said it is going to explore the possibility of further features in the future including providing journey planning information and status alerts. It already deals with a large volume of queries every day through its two Facebook pages.
Shashi Verma, TfL's director of customer experience, said: "Millions of people already use our Journey Planner and social media channels to help them get around London, and we are constantly seeking new ways to make the process even easier. This TravelBot will make it simpler for people using Messenger to get the information they need as they move around the city.
“We think that this initial version will be a major step forward in how we provide travel information to our customers and we look forward to their feedback to help us improve the product over time.”
The move is the latest from TfL in providing new channels for customer information and making its data more widely available. It has been a leader in the open data movement, with more than 600 apps now powered by the data from its APIs.
Image by Aubrey Morandarte, Guildford, CC BY-SA 2.0, via Wikimedia Commons