There’s been a lot of talk about voice and conversational UI at Tonic Design. We recently released Change The Subject in the Amazon Store (check it out, and be sure to use it for those awkward holiday get-togethers), and we’re already working on more voice experiences for Alexa and Google Home.
One of the challenges that comes with designing for voice is how different it is from traditional tech. The user experience relies solely on the conversation; traditional visual elements you’d use to guide the user or to tell a story aren’t applicable to voice.
At Tonic, we have a system we follow to ensure that we can create meaningful conversations for our clients and their customers. Our voice and conversational UI process is divided into two paths: Our developers focus on the technical aspects of making a skill or utterance, and our conversation UX designers are focused entirely on the user and how they’ll interact with what we’re building.
We illustrated Tonic's voice design process on the left-hand side of the post and our development process on the right. By looking at each, you’ll be able to see what both disciplines bring to the project to make a meaningful voice experience for users.
Empathy and understanding the user’s needs are key to designing for conversation. We call this purpose. We need to find purpose to help build conversations in Stage 3.
To get there, we map the customer’s experience as it relates both to their own personal life and to the brand as we gain empathy insights. From there, we can develop use cases (purpose) for building conversations.
While we take time to understand the user, we also learn about the brand. Through market analysis and examining business strategy, we see what objectives we need to reach through voice.
After we map the customer experience, we can start to shape the user persona. The persona should be formed with the right mix of demographic and ethnographic information.
From here, we can begin to shape a brand personality. This brand personality has to be deliberate and apparent. We’ll apply this in Stage 4, where emotional connections are made in refining our conversation.
With the information we gathered in Stage 1, we can now prioritize three use cases for building separate conversations. These are rough conversation “happy paths”. This is a very quick way to plot a conversation from beginning to end without all the added detail.
This is where the conversation begins to take shape. We develop greetings and goodbyes, assess incorrect answers, and dive deep into utterance mapping.
An utterance map accounts for every possible way someone could react or respond to one question (for example: “What time is it? What’s the hour? Can you tell me the time? What’s the time?”).
Once we have the conversation mapped out, we need to ensure that the brand persona shines through. The persona should be consistent throughout all use cases in terms of tone and personality.
Amazon has a variety of speech variants that affect the way it pronounces certain words and syllables. The platform’s use of speechcons, (special words and phrases that Alexa pronounces more expressively) can also add some personality and depth to the experience.
Testing is a critical part of everything we do at Tonic. Real people need to be involved in the testing process to ensure that we’re designing for the user. Developers and testers should work together closely to determine the best way to hand over final conversation and utterance maps for testing. Time listening to the utterances and interacting with the skill is essential; not everything that looks good on paper sounds as good when spoken out loud.
Designing and developing for voice is a unique experience for everyone involved in the process. Our 5-step process works with the current state of voice technology, but as it becomes more advanced, we anticipate adding even more steps to the process to account for new capabilities.
If you want to learn more about what voice and conversational UI services we offer, check out the voice page on our website.