What is the future of bots

digitalization

"Alexa, what's the weather forecast for tomorrow?", "Hey Google, order me a pizza!" For us today, simple inquiries are just a voice command away. We are more and more used to using intelligent language technology to make our everyday lives easier. However, we still hardly scratch the surface of the potential of our new virtual friends. Take, for example, Google Duplex. In this demo, millions of viewers witnessed the abilities of a very human-sounding voice bot firsthand as he made a reservation with a human over the phone. What if these skills evolve and calls to a company can be made entirely by AI-powered technology? Will bots soon be able to act on your behalf and use the bank's bot to control your accounts or transfers?

It's not that far yet. But the possibilities and expectations of voice control have changed. Users no longer have to rely on IVR (Interactive Voice Response) technology and laboriously and repeatedly struggle through individual queries just to obtain an insurance offer, for example. Instead, the concerns are communicated online, understood and answered immediately and precisely. By using natural language processing and speech recognition, bots are now able to understand human conversations and process them according to a set of rules.

The easy way

The fact is, it's so much easier to do research than reading dozens of pages of text, using unfamiliar software, or searching, scrolling, and typing for a long time. With advanced language technology, we get what we want, when we want it. Speech works intuitively and without going through input devices. In short: it is simple.

Google says its technology can handle complex conversations without the need for a human authority. Even if Google has not positioned the technology as a business technology, it is very likely that it will also be used in business life - especially if companies want to use voice interfaces. Primary application scenarios are organizational functions on workstations and networks, such as voice-controlled search or digital assistants such as Microsoft Cortana.

However, it is important that the focus of the language technology is on the benefits of the integration - not the costs. Used correctly, consumers will appreciate voice bots in the corporate context as much as they do at home. Provided that they are just as effective, available and responsive. The (operating) comfort is crucial. Nobody likes to call the bank and then wait five minutes on hold, only to speak to a completely unprepared employee. If voice bots add value, consumers won't hesitate to use them.

The big picture matters

But comfort isn't the only factor. Both companies and users need to be comfortable with the technology. And there are numerous basic social, moral, and legal implications to consider in order to find and maintain that balance. How do you make sure that a voice bot is ethical? How do you prevent the use of inflection and sentiment analysis to manipulate people during a bot conversation? What do bots do with the information that users provide? Will they remember credit card numbers? Where does the information go? How are they saved? Who else can access it? In Europe, for example, voice bots have to comply with the EU GDPR, which came into force in 2018 - how can this be guaranteed?

Human-like voice bots require a large amount of data to operate and providing this is associated with an enormous responsibility. Companies need to know where their data comes from and where it converges. Are internal data combined with data from third parties? Where does this data come from, and can and should you use it for your own AI models? If past scandals and data leaks have taught us anything, it is that companies do not always correctly assess and take into account the negative effects of popular new technologies and their potential for abuse. That requires planning and due diligence.

Companies that only collect and process their own data are therefore particularly suitable for getting started with this technology. Here you can create an isolated test environment and check the practicability of voice control in your own company context without endangering data security. If the feedback is positive, the technology can then be gradually introduced in other departments and more complex data networks - but always with a view of the big picture of your own IT architecture.

Meet user expectations

In addition to legal and ethical hurdles, companies should also be prepared for the high level of customer expectations when using voice bots. As humans, we are biologically programmed to recognize voices and to instinctively remember our previous interactions and actions with the associated persona. If a company's voice bot sounds like Alexa, users will also expect it to "behave" like Alexa. However, if there are discrepancies with user expectation, it will quickly lead to annoyance among consumers.

Conclusion

Even if we encounter voice bots everywhere soon, we will not always want to talk to them. They will be particularly useful for quick queries and simple tasks. But the feeling that arises when talking to a real person about complicated, sometimes very personal matters, can never be fully achieved with a bot. The complexity of human communication and the associated sense of authenticity and our unique ability to share experiences with one another - this is a type of voice interaction that can never be synthesized and automated in a fully satisfactory manner.