Voice bots keep getting better and better, but their growing adoption raises at the same time questions about data protection and the limits of use in personal conversations. Although, we are clearly only at the beginning when it comes to using intelligent language technologies, their integration into the customer dialogue is constantly progressing.
Google CEO, Sundar Pichai, for instance, played once at the Google I/O two recordings of Google Duplex conversations. The assistance system arranged initially confidently an appointment with the hairdresser, but when it came to the second request to reserve a table for four people in a restaurant, the virtual assistant failed because it was only possible to order a table for five or more people. After all, Google Duplex was able to elicit from the restaurant employee that there is a prospect of getting four seats at the planned time of the restaurant visit. In both phone calls, the people called did not notice that they were talking to a machine.
The Google Duplex demo shows what AI-based technology can do: it can, under certain circumstances, have complex conversations without the need for a human to intervene.
And if we reflect on this development, it will not be long before users can instruct their intelligent virtual assistant to call the bank to clear some transactions with the bank's bot. Voice-controlled searches in computer networks could also be other useful applications for voice bots.
Be that as it may, as voice control develops, it also creates expectations. Voice-activated devices like Alexa, Cortana, Siri, and Google Voice are ubiquitous and are being used by more and more consumers.
The younger generation is becoming more proficient in using language technology to get answers to their questions. These users simply no longer want to interact with Interactive Voice Response (IVR) tediously and repeatedly until their request is resolved. The trend is therefore clearly going into the direction of communication via voice and this is exactly where voice bots will come into play. A well-trained voice bot can handle certain scenarios itself. The system then processes the call according to a set of rules. If, at the end of the day, the added value comes in the form of quick and precise information, consumers will not hesitate to use voice bots.
Availability, effectiveness as well as speed and responsiveness are the criteria by which the new technology stands out from contact centers that use outdated IVR technologies. When businesses use voice bots, it should be done in combination with human agents. Because when more empathy is required, human employees can expand the service bot's offering to meet customer expectations.
Not only the users, but also the companies must feel comfortable with the technology. In advance, those responsible should therefore play through the basic social, moral, and legal consequences that can occur. They are faced with the task of finding a sensible balance between the various factors, to which the examination of the following questions can contribute: What must be done to ensure that a voice bot behaves ethically? How do you stop inflection and mood analysis that people manipulate in a bot conversation? What do bots do with the information they get from users? Does the machine forget the credit card number immediately? Where and how is customer information stored? Who can access it?
There are also other questions that need to be clarified, such as: Are internal data combined with those from third-party providers? How did the third-party provider collect this data, and can it be used for the own AI models?
Past scandals and data leaks should remind us of how important these aspects are. Here, companies got into the headlines because they misjudged the negative effects of popular new technologies and their potential for abuse or did not consider them at all.
Voice bots will seize communication in customer service where quick queries and simple tasks arise. However, human language and how to use it is complex, which is why it can quickly overwhelm a virtual language assistant and the conversation becomes too personal. Authentic appearance and sharing of experiences can never be synthesized and automated in a completely satisfactory manner. This means that person-to-person contact remains an immensely important component in customer communication.