Your Voice Design Interview
How to “talk the talk” when interviewing for a voice user interface design role.
An increasingly common inquiry I receive these days is from hopeful first-time voice designers facing an interview. “What should I expect?” “How can I prepare?” “Any tips for interviewing for the Alexa team?” Rather than dole this advice out one-on-one via messages, I’m taking to Medium to share a few thoughts.
Of course, before you pursue a voice design role, I hope you’ll reflect upon whether such a role at this stage in your career will truly provide satisfaction and growth. A dedicated role in voice design isn’t for everyone, especially if you’re visually motivated. You may also want to read my earlier thoughts on this career leap, “So You Want to Be a Voice Designer”.
But for those of you further down the path, facing an impending interview: read on. Three steps to steel yourself for those tough conversations and design exercises as a first-time or early-career voice designer.
Step 1: Immerse yourself
If you’re several weeks out from your potential interview, or even if you have just a long flight to work with, my top recommendation is to take some quality time immersing yourself in the few books out there on the subject of voice design. A great place to start is my earlier Medium post on reading recommendations. If you must choose only one at this time, go with Cathy Pearl’s overview: it will give you the broadest sense of the specialty.
If you don’t have access to books in the short term (or if you’ve already tackled them), spend some quality time with the skills ecosystem of your favorite voice assistant. Install some 5-star skills…and some one-star skills. It’s easy to take first-party voice UI decisions for granted until you see the other side of the coin. (Mind you, truly terrible VUI doesn’t get through certification, but there’s still a wide range of design decisions evident in the VUI skill community.)
Step 2: Scrub your portfolio
If you are interviewing for a voice design role without prior voice design experience, don’t despair right away. There are some parts of your process and deliverables that will translate more directly.
Avoid presenting a portfolio that is simply a set of screenshots, even interim ones. Your potential team will need greater insight into how you think, and how you manipulate abstract concepts that do not have visual equivalents. To that end, I recommend reprioritizing your portfolio presentation to highlight the following types of work:
The ability to arrange abstract concepts in a way that reflects your customer’s mental model is relevant to all forms of interaction design. Any projects that lean heavily into information architecture are good candidates for discussion.
A critical voice deliverable for my engagements was interaction flows. Representing the various decisions, inputs, and outputs in an end-to-end fashion helped my development teams bring our voice designs to life in the most robust fashion. If any of your work on visual interfaces includes complex interaction flows, consider emphasizing that work.
User research & data-driven design
Because mainstream voice design is still generally in its infancy, we still have a great deal to learn. Voice designers must be humble, curious, and agile in their approach to dealing with customer feedback, emergent behavior, and telemetry. Highlight any portfolio projects that show non-trivial adaptations based on customer feedback or data mid-project. These skills are universal, but perhaps even more critical for voice designers.
While I believe all of the major voice assistants are supported by dedicated content teams, my experience was that those resources were over-tasked. Voice designers were the first line of defense, expected to write reasonably crisp, on-brand content. We only escalated for extreme cases. However, the key word here is crisp: you won’t gain points from a voice design team for overly flowery, high-word-count prose. This is the hardest competency to convey in the context of a design portfolio review… but if you have an applicable example, include it.
You’ll inevitably only have time to highlight a few projects, so choose the projects that reflect some of the above points most strongly. As always, a quality walkthrough of a project end-to-end, with interim deliverables, is always a stronger proposition than a large quantity of poorly documented examples.
Step 3: Reframe your thinking
Inevitably, you’ll be asked to participate in a design exercise. If the role you’re interviewing for is truly voice-only, it will be important to show that you approach voice designs with a different perspective. The way you approach this design exercise will provide your future teammates critical information about your growth mindset and flexibility.
Some questions you may want to ask as you frame your voice design problem:
- Where will the customer be using this interface? (Living room? Kitchen? Noisy public space?)
- With what type of hardware will the voice experience be delivered? Are the microphones far-field (Echo) or near-field (phone, Tap)?
- Will other customers be present during the interaction? (Voice interfaces have a performative element when used in public.)
There’s no way to teach you voice design in a single article (if you’re interested in going deeper, I do offer full-day VUI workshops through Ideaplatz). However, in the short term, remember these things:
VUI information architecture is broad, not deep.
The last thing a voice customer wants is a deeply nested series of prompts. One of the greatest strengths of a voice UI is how it enables breadth of access. You’re not limited to the rule of 7; there is no risk of visually overwhelming a customer. As long as potential prompts and intents match common mental models, your customer can achieve many intents from the same neutral starting point.
Advanced VUI often requires manipulation of variable content.
Think about the most common and most loved abilities of your voice assistant of choice. Odds are, you set timers, set alarms, and play media content. What all of these interactions have in common is that they depend upon variable content — called “slots”. In the utterance “Set a timer for 30 minutes”, the value “30 minutes” is probably represented in the design by a duration slot. Early voice design requires identification of the full range of slots customers will attempt to use with an intent (the customer’s desired action). Make sure you identify these in your design exercise, and if they are unclear, ask questions to get clarity.
Voice UI is largely about error correction.
The things we say to our voice assistants are often misheard — whether because the customer phrased a request informally, whether ambient noise interfered with recognition, or whether another unforseen problem arose. Where most GUI design exercises focus primarily on a “happy path,” VUI designs spend comparatively little time on the desired outcome. Most of the effort is dealing with empty slots, malformed or invalid slot values, or perhaps complete misrecognition. Don’t forget to account for when things go wrong.
Don’t be tremendously discouraged if you encounter rejection at first. Rejection is always painful, but we all encounter rejection at some point. My first interview with a natural user interface team was the Kinect for Windows team back in 2012. At the time, I was told I didn’t have enough experience with voice and gesture. I was very frustrated, but I simply continued to seek out opportunities to grow. Looking back, that certainly seems like an ironic moment, but it was my determination and curiosity that helped me turn the tables in the face of that rejection.
The good news is that even if you’re rejected, your options are more numerous than mine were at the time. There are more books, online guidelines, and plenty of Medium articles. There are numerous talks and workshops (including my own.) And, of course, there are the skills kits! You are only a few hundred lines of code and a few developer accounts away from building, and perhaps even shipping, your own voice experiences. I hope to write more about my Skills experiences soon.
In the meantime, good luck! May the voice be with you.
Cheryl Platz has worked on a variety of voice user interfaces including the Echo Look and Echo Show, Amazon’s Alexa platform, Windows Automotive, and Cortana. She is currently Design Lead for the C+E Admin Experience team at Microsoft. As founder of design education company Ideaplatz, Cheryl is also touring worldwide with her acclaimed natural user interface talks and workshops. In the first half of 2018, you can join her at Interaction ’18 and UX London.