How can chat bots be used professionally? After our unscientific self-experiment with the Replika app, we wanted to learn more. That’s why we took a closer look at practical use cases of digital service agents. The first part of our analysis is dedicated to chat bots in customer service.
Who still remembers the annoying bug-eyed paper clip “Clippy”, who made Microsoft Word since 1997 unsafe with well-intentioned advice? “It looks like you want to write a letter…”, Karl Klammer (that’s the German name) started, before offering his help without being asked. No thanks, we’ll manage! Whoever is still recovering from a Karl Klammer trauma must be strong now: Clippy is coming back – but only as an emoji.
Chat bots, however, are here to stay. Gone are the days when programmed chat assistants were just a nuisance instead of a real help. Once set up, the bots now do their job without complaint. E-commerce has been accounting for a growing share of retail sales – and not just since Corona. It’s clear that customer service on the Internet is also in demand.
The Internet is open around the clock. There is no lunch break or night’s rest here. Customers expect fast answers and constant availability from competent service. Whether it’s spontaneous advice on a purchase, urgent IT help or quick access to a digital manual – an automated online customer service must deliver exactly what a classic help desk ideally offers. This is why providers of Botcore chatbots are attempting to cover these diverse services.
This calls for solutions that do not involve any additional effort for the customer compared to high-quality human customer service. Ideally, time can also be saved on the customer’s side. Classic models in automated service frequently outsource the time problem to the customer. Everyone knows how frustrating it can be to be stuck in a telephone waiting loop, being asked for “a little patience” every minute, while “the next free employee” cannot be found far and wide.
Systems that allow the user to navigate in a rudimentary way using simple voice commands (“…then say NO.”) or telephone keys (“…then press ONE.”) are one step ahead. However, both models usually provide only limited satisfactory communication – the given paths do not always lead in a straight line and sometimes do not lead to one’ s goal at all. The consequence: In the end, a flesh-and-blood employee has to process the request after all. Competently programmed bots promise a remedy here. Unlike a human employee, they can handle multiple requests at the same time. Simply put, chat bots are scalable. The eye of the needle that determines capacity will then no longer be the staffing key, but the server capacity.
Another problem with the old bots: they are quite limited and do not learn. Modern bots are advertised as artificial intelligence – certainly a bit of a stretch, even if the digital service workers do in fact develop their algorithm automatically via machine learning. In fact, it is now possible to have bots handle even complex questions. Specialized providers such as Crealog promise that their systems not only learn, but also rely on AI support in dialog to better understand the wishes of the human on the other side. Bots appear in this context not only as writing in the Messenger window, but also in the form of “voice bots” on the phone. Unlike the aforementioned tape announcements or yes-no computers, both forms can respond to the counterpart and offer solutions situationally. This makes flexible and therefore goal-oriented communication possible.
Installing a bot does not mean discontinuing human customer service. Online printing in particular is a highly complex industry that thrives not only on mass production, but also on individual solutions for heartfelt projects. In these cases, empathetic advice is needed that thinks outside the box and away from prefabricated solutions. Many projects are about personal feelings, wishes and dreams – even the best chat bot can’t reflect that. The future therefore belongs to the division of labor: The bots take over routine inquiries, technical queries and “frequently asked questions”. Humans will remain important as correctors and vision accomplices. And for retired bots like Clippy, there is still an emoji memorial that reminds us of the bad old days.