Voice Design

Persona design for voice assistants

Voice assistants should not only answer questions quickly and in a satisfactory way, they should be pleasant to communicate with and cope with the curiosities of human language. The development of a persona ensures that users like to speak to their language assistant over and over again. In addition to vocal characteristics, choice of words and sentence structure, intercultural and language-specific parameters must also be included in the design of a coherent persona. You can check out on our website how we did this on our project for “Hey Swisscom”.


Successful communication makes you happy – who likes to fall closed ears? We don’t like to talk to someone who doesn’t really understand, who is empathically on a different level than us, in short: who doesn’t get us.

We humans tend to humanize things of all kinds. Voice assistants are no exception here, on the contrary: speaking is our most primitive way of communicating. When we speak to somebody, we attribute human qualities. In communication, we expect more than just a reaction from our counterpart. We want to be understood, both in terms of content and in human terms. If we feel that someone does not react adequately, we first react critically. And, probably we don’t want to invite them in our home.

Voice assistants live with us at home or even in our pockets and have now become everyday conversational partners. So the persona of a language assistant is extremely important for our acceptance. Anyone who tries Alexa in different languages can quickly notice that – depending on the cultural and linguistic context – it affects us completely differently. It’s not a coincidence.

A well-known principle from communication is: “You can’t not communicate.” This includes body language, gestures and facial expressions. But the tone of the voice and the speaking behavior also seduce interpretations: what is the meaning of what you have just heard? These interpretations are mostly passive. Whether we want to or not, we automatically identify the speaker’s character on the basis of hearing. And with that it is decided whether we want to stay in contact with the other person. For example, who likes to hear one call after another or wants to be overwhelmed by questions? But sometimes it is only very subtle nuances in the speech design that decide when something is interpreted as a request or a question. It all depends on who is listening and who is speaking.

That is why the design and configuration of a persona in the development of voice assistants is an essential success factor for subsequent acceptance. First of all, the target group needs to be clarified.

Which users will talk to the voice assistant?

What kind of world do you live in?

What kind of speaking behavior do they maintain among each other?

The more specifically the users are recorded, the easier it is to identify the persona as a communication partner:

So what are the characteristics of the persona as a conversational partner?

What will your voice sound like?

Will she speak at eye level, what words will she use?

Share via