Last update: 1 Sep, 2020
Thanks to digitalisation and rising customer expectations, the number of customer contact points outside the service department is increasing in virtually all companies. At the same time, work across departments is strongly characterised by transparency, performance and a high customer experience - or at least the desire for it. AI-based bots and smart assistants promise support. But even with the simplest of enquiries, a bot is not automatically the best solution - the keyword is tact. Really smart solutions can still help.
Customers are particularly dissatisfied when companies have problems even with standard enquiries, when it takes a long time to process them or when customers need several attempts to resolve a minor issue. Experience shows that standard enquiries account for about 50 to 60 per cent of all enquiries to companies. So there is plenty of opportunity for failure. The fact that even answering simple requests sometimes leads to disappointment is due to the fact that employees are confronted with a wide variety of contact channels and data-carrying systems throughout the company, which makes the search for information more difficult, especially under time pressure. When it comes to complex processes, the problem is magnified. Customers are often left bewildered and frustrated.
AI-based bots promise support. They automate telephone dialogues and the processing of non-voice contacts partially or fully automatically. For example, voicebots can relieve employees of standard telephone questions, process chat messages on a case-by-case basis, pre-qualify and forward messages or change master data. A really smart voicebot receives telephone calls and understands the request through powerful speech recognition and speech analysis. This means that customers have to wait less and their concerns are resolved quickly. The same applies to email and chatbots. When done really intelligently, bots push the customer experience noticeably.
Mood- and intent-dependent routing can distinguish simple, repetitive requests from complex concerns, or well-tempered customers from angry quitters, route a call to an investment advisor or a self-service. This is decided according to the complexity, urgency and mood of the customer. By asking an open question at the beginning of the call ("What can we do for you?"), the AI recognises the intention of the caller and routes demanding enquiries such as the financing of a property to an investment advisor or, in the case of simple concerns such as the account balance, to a self-service. The self-service agent handles the request promptly, error-free and around the clock. Great.
But even for the simplest of queries, a bot is not automatically the best solution. "I have a question about the order status" indicates a relaxed, simple process. "I want to know where my order is!!!!!", on the other hand, indicates annoyance and a disturbed process. Connecting the customer with a self-service bot would be possible on an intent-based basis, but not very clever. Ideally, the smart sentiment recognition would then refine the routing and head for the employees.
And then there are the smart assistants that support employees in their work. They take notes of relevant conversation content and make it available immediately - taking notes is a thing of the past. This not only helps the staff to concentrate, but also avoids mistakes. Some buddies transcribe the recognised content of an enquiry immediately into data-carrying systems and make the relevant information available. And if further process steps are necessary after processing, such as creating a file, informing specialists or making an appointment, the assistant takes care of this. Such systems considerably relieve the customer service and push the customer experience without replacing humans.
The important thing is: valuable employees should not work like a bot, but take care of complex, demanding tasks. Conversely, a bot is not a human being and its abilities are limited. Whenever empathy and reflection are required, when it comes to calming down angry customers, finding errors or correcting automated decisions, a human is needed. A bot can't win back cancellations or cross-sell. The idea of complete automation will probably lead to a customer experience in the dumps and possibly alienate customers much more than an overworked employee, with whom customers can at least vent their frustrations. Even if this is not nice for the staff. Customer experience is and remains a human thing, because bots don't order shoes or a new car and AI has no clientele. The decisive phases of the customer experience should not necessarily be left to a bot. But if we use AI as active support, suddenly many things become possible!
Author: Andreas Klück