The most useful secret to making a good chatbot or voice assistant
Conversational AI is a thing. It's being used for chatbots, voice assistants, and all kinds of robots. Designers nowadays have to design conversations for all these different interfaces. But how do you design for an interface where people can say and do whatever they want... whenever they want?
How would you go about designing for multiple channels and interfaces? Does that not cause us to do a lot of double work, and how complicated is this? These are the main questions that we get from companies and designers from all over the world. Everyone is trying to figure out how to effectively design conversations. In this article, we share the most useful secret to making a good chatbot or voice assistant for multiple channels.
The secret is to design voice-first
What we mean by that is to focus on the bare minimum. If you can have a conversation in a dark room using just your voice, then it works everywhere. It’s the hardest and most advanced design skill out there. You have to work without any type of visual cues. No text, no images, no buttons. Speech only. If you can design for voice, you can do it all.
When you focus on the bare minimum of voice design, and you would like to utilize that design for a chatbot, you simply add buttons to make it easier for users to navigate. But if you start by designing a text-based chatbot, you incorporate buttons from the start. Then scaling your operation to a voice experience like Alexa or Google Home would mean having to design everything all over again.
Designing voice-first also ensures a more natural conversation. The best way to design for voice is to start with sample dialogues. It is a kind of improv theater. You put two people back-to-back. One person plays the user and the other plays the bot, and you let these people have the conversation for you. Record and document the conversation and you will have a solid basis for designing a natural conversation.
The next big thing is multi-modular interfaces. A combination of voice, chat and screens. They are becoming increasingly popular. Your voice-first design will come in handy here. Because if you designed your conversations to work for voice, you can now add visual cues to enrich the experience.