Due to voice assistants that are used in customer service, such as Siri or Alexa and “talking robots”, the topic of speech recognition is currently highly topical. Voice assistants learn to decipher the core of a customer’s message, so that the customer does not have to learn which commands to use to get to the goal, which sounds easier than it is, considering that dialects, synonyms or paraphrases have to be learned by the bot first.
Incidentally, speech recognition does not only apply to spoken, but written language, too. In fact, the chatbot is currently the best-known form of artificial intelligence (AI) in customer service, since the text-based dialog system can handle simple inquiries and tasks and has the advantage that it is available 24/7.
The more popular chatbots become, the more it is necessary that they can respond to different requests in such a way that the customer does not have to constantly rethink it until he has found the right command. For customer service, however, the chatbot is just one of many AI applications that can be used to optimize the customer experience – and by no means it is a substitute for live chat. Rather, it can be a first point of contact for customers outside of working hours or to solve simple inquiries without the need for a service employee. This relieves both customers and employees by avoiding putting someone on hold. Not to mention that especially in the form of apps and personalized customer accounts, self-service can have significant advantages for customers when it comes to repetitive, simple processes. Besides, based on data collected from the customer and data from similar customers, it is possible to determine the probabilities of how successful the respective measure will be. In combination, you can contact customers long before they decide to bail out of a purchase or bind them to the company with the right offer or the right strategy.
Certainly, not every customer inquiry has to be dealt with immediately, but there are requests for which every second counts to prevent a bail out or a negative experience. With the help of AI, the ticket system can learn which requests (from which customers) need to be classified as more urgent than others. The employees in this case do not work chronologically and thus risk that an important case must wait longer but can work on the cases according to priority and urgency.
With the use of AI in customer service, companies aim to achieve an optimized customer experience across channels in real time. Above all, AI technologies are an important tool for merging different touchpoints quickly and easily in an integrative approach.
Application scenarios that can be implemented in customer service using AI and machine learning methods include:
Realtime ‘next best action’: based on predictive analysis and machine learning, AI engines can automatically suggest the “next best action”. An optimal solution not only evaluates historical data, but also takes into account contextual information that arises at the moment of the specific interaction - for example the reason for a customer call to the call center or the time during which he was on hold. Since AI predicts the current needs of each individual customer across channels, it can make a decisive contribution to cross-selling or up-selling. It is important that the system reacts in real time: if the customer rejects the cross-selling offer, for example, the system should immediately submit an alternative proposal.
Automated dialogues: AI-based text analytics helps answering customer questions in chat rooms or email. With chatbots, i.e. through dialogues automated with algorithms, answers can be found automatically for a large number of customer questions - possibly even better than those provided by the call center staff. However, clearly defined tasks should be defined for chatbots; in principle, they are most efficient when they address a defined topic.
Sentiment analysis: AI can help to assess the mood of the customer, for example whether he is angry or in the mood to buy. The AI-based sentiment analysis, for example, can be used excellently in the field of chatbots. It can be used to identify whether a customer has problems interacting in the chat. This also makes it possible to determine the optimum point in time at which a call center employee should take over communication.
Predictive customer service: AI can help identify problems and concerns of customers before they even contact customer service - for example, a DSL customer who has not had an Internet connection for several hours is very likely to contact technical customer service in a timely manner.
Intelligent routing: AI can also help optimize the distribution of customer inquiries in the call center and customer service and helps in the classification of the concern and the identification of the most appropriate service employee - for example in terms of expertise or availability.
However, a typical obstacle to a consistently positive customer experience is customer data that gets stuck in data silos. Here AI can help to clear the complexity of the data or measure customer interactions, recognize behavioral patterns, and then evaluate them. The results obtained can then be made available to employees in marketing, sales, or service across all communication channels. This enables a 360-degree view of the customers and provides them with a basis for personalized offers with a positive effect on the customer experience.
But one thing remains certain: a high-quality and personalized customer experience is not possible without human involvement. Against this background, the path to success for companies is not to replace employees with AI, but to train them in such a way that they understand their own key role within the customer journey.
By Daniela La Marca