AI in Customer Service: Success Factors

Think back to the last time you had a problem with a service provider and had to telephone or go online for help. Did you need to update your address, query an incorrect bill, ask for an upgrade, or just ask a question? Whatever the issue, I’m sure it wasn’t a pleasant experience. No one likes calling up an automated phone line, being passed on to an operator who asks you the same questions you’ve already answered then only to transfer you to another department to start over again.

Deutsche Telekom have identified this problem and Miles Lynman-Smith, Chief Programme Manager, explained at the Deep Learning Summit in Montreal that ‘no one enjoys speaking to a telco as the experience is repetitive long and often inaccurate. This was also true for the agents, they didn’t like doing the same job over and over again - they wanted to spend their time on solving complex customer problems rather than trivial issues.’

Having identified the problem, the team at Deutsche Telekom knew that an AI was the answer, but they were faced with much skepticism from their customers. People are scared of the unknown, and AI is largely still unknown. Miles explained that 'people have this fear that we’re going to be turned into robots, or have our jobs stolen by AI: people were scared of AI because they didn’t understand what it actually is or what it’s there for', so Deutsche Telekom set about defining it for their customers, and they chose to define their AI as something that can ‘listen, learn, optimise and act.’

At the Deep Learning Summit in Montreal earlier this Autumn, Miles shared his work on their AI assistant and explained how they managed to transform the customer opinion of AI so that it was viewed as a customer service assistant, not as something to be feared. The approach was simple, replace the repetitive everyday boring requests using cognitive AI. This however, still posed numerous concerns such as job loss within Deutsche Telekom, lack of customer satisfaction, etc). In his presentation, Miles explained how they overcame these issues and expanded on their next steps towards bringing AI from an innovation on the periphery of the company to an innovation in the centre – automation of processes will totally change the way we work.

Watch Miles’ presentation here to learn more about launching AI in enterprise.

Watch more from the Deep Learning Summit in Montreal on the video hub and hear about the latest progressions and industry applications of DL and AI.