Voice User Interfaces (VUI) are a development that companies should urgently deal with today.
“Alexa, ask Deutsche Bahn for a connection from Hamburg to Frankfurt!”
Can you imagine booking a train like that? Or would you personally prefer to ask Siri or the Google Assistant to visit the website and book the connection yourself, make the booking via ChatBot in WhatsApp or Facebook Messenger, access an app, or talk to an employee at the train station?
No matter what your personal preferences are, the decision for the user experience optimized to the customer situation and the subsequent implementation present every company with enormous challenges.
Customer expectation of a Voice User Interface VUI
Your customers expect to be offered the simplest and most preferred communication channel for them. And true market leaders are characterised by being the first to exploit the advantages of the individual channels and thus offer customers real alternatives to competition. The customer decides within 30 seconds at the first attempt whether he will use this channel a second time. This offers very little space for experiments.
Differences to traditional user interfaces
There are striking differences in the characteristics between the individual forms of communication. Voice input must be completely different from input via a web page or within a messenger, as the visual information is completely omitted. The development of a voice user interface makes completely new demands – it is much more serial and also much slower than a graphical user interface (GUI). The ear can only process one piece of information at a time, while the eye recognizes patterns from different symbols in fractions of a second and thus processes multiple information simultaneously. The entire language-guided input process is slower, breaks have to be included, inputs have to be confirmed and verified over and over again. This achieves a high degree of clarity and understanding on both sides. In addition, a certain degree of sensitivity is strived for, which gives the customer the feeling of personal care.
The customer has very specific expectations according to his situation and quickly forms an opinion about the quality of the dialogue, which is comparatively slow. Dialogue management must therefore get to the point very quickly and deliver selected, suitable results. In our example of train booking, the possible selection is reduced to two or three meaningful connections, more can only be represented with difficulty – the requirement for the relevance of the results is therefore much higher than with the booking process on the website. In this respect, the previous dialog should have requested additional information in advance, know customer preferences or have additional restriction algorithms.
The more complex the business process, the more complicated the development of this dialogue becomes. The company must therefore first develop a dialogue model that is completely independent of the previous models and optimized for voice input.
Data is the most important quality feature
All forms of communication have in common the quality of the underlying data as a quality feature. In the past, the customer was upset about the call center employee, who had no access to the details. Today, he expects the bot to know not only the basic data, but also his own preferences or additional data such as his location or current events. The customer assumes that the bot recognizes where the customer is, knows the current traffic situation and also takes into account that bad weather could have led to train delays and therefore makes a completely different travel proposal.
All this requires consistent planning at both the brand and technology level to lay the foundation for execution at the product level. In addition to the conditions mentioned above, the application scenarios in particular must be considered. In which situation which customer decides for which communication channel, and what does he expect in this particular case? When is a voice first approach that offers the customer an alternative to a website or a ChatBot in a messenger useful? How large is my target group for VUI and what are its preferences? Which mechanisms can I use to clearly identify the customer so that I can avoid the recurring entry of customer data and still guarantee the necessary data security? And finally, the question must also be clarified as to how high the development and implementation costs will be compared to the expected return and when the development will be sustained.
First develop the brand strategy for Voice and VUI
Marketing managers and CMOs are currently bombarded by various providers with information and suggestions on how they can quickly and easily implement these new communication channels in initial test projects. We recommend that you first deal with a holistic planning and subordinate the level of tools and platforms. Instead, the technologies in the areas of AI, language and integration that are realistically applicable today are determined and a realistic estimate of the development is made in the coming years. The technical requirements and the roadmap define the providers that will have an impact at the implementation level. In this environment, the selection of one or more platforms is a long-term decision that cannot easily be tipped over again. Current use cases, the future roadmap and the explicit proof of competencies are important to identify the optimal platform and the right service provider.
In the end only one thing applies: the customer will no longer forgive you for any mistakes – you have little opportunity to experiment. You have 30 seconds to convince him of your communication.
This article was also published here at ISG Research.
Translated with www.DeepL.com/Translator