Voice User Interfaces and How to Write for Them
Mar 6, 2017
Job fairs get busy. Typically, long lines of well-dressed candidates form at Centerline’s table, asking, “Do you really have a bar?” and, “What hours would I work?”
Emily Gamiel, recruiting and onboarding specialist, is often asked questions just like these. As Centerline’s full-time recruiter, they’re “softballs.” She could respond in her sleep. Others, though, are more tricky. “Hardball” issues (ex. work/life balance or compensation) deserve individualized treatment. Her discourse around these issues should be the result of experience, intuition, and empathy—all uniquely human traits.
What if we could answer the “softball” questions before job seekers reached the front of the line? Then Emily could focus on helping candidates with the “hardball” issues, like how to get hired and if Centerline is right for their career, instead of spending time on the “softball” questions.
We could hand out a written F.A.Q. But they’re cumbersome and against the personalized spirit of our hiring process.
The Centerline Recruiter Bot
Instead, we’ll “start the conversation” with job fair attendees, by letting them have a conversation…with a chatbot. We’re developing a voice user interface (VUI) that will answer basic questions about Centerline. Once complete, our “Siri-like” bot will join Emily at job fairs, where it will listen and respond to candidate’s questions.
So, what are VUI’s?
They’re close “cousins” to graphic user interfaces (GUI). Instead of (or in tandem with) text, visuals, and interactive forms, VUI’s listen for a user’s spoken input. Compared to GUI’s, they’re more intuitive to use and faster to interact with.
Before you started asking Alexa to reorder your non-GMO, cage-free laundry detergent, it’s likely you had already used a VUI. In the early 2000’s, hundreds of businesses deployed interactive voice response systems (IVR). Typically taking on roles of customer service and support, IVRs were the first computers capable of understanding human speech over the telephone and carrying out basic tasks. The technology was new and painfully awkward – too often, the systems failed to correctly interpret spoken input. And those failures were renowned.
But, advances in computing power and sophistication have improved the accuracy of VUIs. Today, most Americans own a smartphone with a voice-activated personal assistant. In addition, a growing number of households have a voice-only interface, like the Amazon Echo or Google Home. The VUI is quickly becoming a part of daily life.
Marketers who want to build a VUI for their brand can use any of the many free and easy-to-use tools that have recently been made available. Sure, these tools save you the hassle of hiring an in-house data and computer science team. But, before you start scripting the next HAL 9000, there are issues to consider that are unique to VUIs. In the case of Centerline’s recruiting bot, here are just a few:
- How will candidates discover what they can ask the bot?
- Since the job fair attendees will start the conversation, how will the bot handle the variety of possible entry points?
- Candidates may have a variety of questions. How will the bot seamlessly engage in a nonlinear conversation?
- Candidates may ask us questions requiring a lengthy response. How do we break it up into easy to listen-to-and-understand bites? How do we list all the alcohol we have at our bar, without listing it all at once?
- When eager job seekers ask the bot questions best left to the humans, how will it conduct the handoff? How will it handle questions it doesn’t understand or know the answer?
In our next post, we’ll unpack these common VUI issues and how to address them. Have any questions in the meantime? Feel free to reach out on Twitter – I’m @JaredBrickman.