EUROPE explains : “ The chat bot follows the same decision tree as the tech guys in the service centre do , so what we have learned in the service centre regarding the best ways to respond we have programmed into the chat bot .”
“ If the bot cannot solve your problem , it will advise you to call the service centre , “ Bauersch adds . “ For me the next logical step is to take the chat bot conversation and continue this with an agent conversation in webchat . You don ’ t have to switch the channel , the agent simply takes over from the bot .”
Rise of the machines We end our conversation discussing how much integration of AI can and should take place over the coming year . For Bauersch , a key target is the early stages of conversations between service agents and customers .
“ For instance , the first minute of the conversation is always the same – how can I help you ? Do you need help with mobile or fixed line ? May I have your customer ID , address and date of birth ? We believe this part of the conversation can be handled efficiently by a bot .” Another potential avenue for a pilot bot to thrive , with Deutsche Telekom about to trial on Facebook and Twitter an automated feedback process . If the customer reveals to the robot after the dialogue that their question was not answered , they will be reconnected to the agent .
The level of success regarding implementation of AI is all about finding the appropriate balance , as Weisenburger states : “ We don ’ t believe that technology will replace agents completely within the next years , rather it will help them to give better responses .” Indeed , Niemeyer adds that “ technology is not yet advanced to a stage where you can replace whole processes . It is more about which part of the interaction can be supported by a bot .”
Fast forward to 2020 and the end of Deutsche Telekom ’ s five-year customer service transformation programme , and it will be fascinating to see what this balance looks like . www . telekom . com 107