As chatbots, conversational interfaces and other voice-activated assistants are becoming more commonplace, the notion of “conversational design” is an emerging space for UX designers.
While we may communicate one way conversationally with another person, communication with a machine brings up different challenges and nuances.
One way to improve the conversational interface design is to develop a strong identity and persona for the voice.
Brad Abrams, group product manager for the Google Assistant Platform, revealed the conversational bots with the best retention also had the strongest personas.
Just as a company’s brand is strongest by delivering on its promises and mission, a well designed voice interface has a consistent identity and persona via voice. Any inconsistencies or breakdown in response creates mistrust and frustration for the end user.
Developing a Conversational Persona
Here are some tips on developing a persona for your conversational interface:
Consider the context: Adapt your voice to the goal/role of the interface, use key scenarios to determine what level of complexity or tone may be appropriate.
Make the unhappy/repair path awesome: In the case that your user responds using a phrase that your interface can’t answer, optimize for a smooth entry back into an area where the bot can assist.
Be natural: Follow basic conversational flows like a greeting or randomization of common messages. Repetition of common phrases can quickly feel stale and automated for a user. Don’t treat users like machines and expect all users to answer in the same way.
Keep it simple and short: No unnatural jargon. Use transparent, crisp and clear language.
Use supplementary sounds: As most people still don’t feel that speaking with a conversational interface is intuitive, designers can use computer generated sounds known as “earcons”, to indicate state and move a conversation forward.
Though conversational interfaces may feel new, previous design guidelines and principles do still apply. Having a basic user-centered design mindset is relevant to conversational interface design, which means creating a system which supports its intended users’ existing beliefs, attitudes, and behaviors. Finally, make sure to go through your flow with real users to get an understanding of where there may be areas for improvement.