Select Page
Highlights from State of Digital Quality

5 Lessons Learned from VoiceTalk: The State of Voice for Enterprises

There’s no question that voice is the next digital frontier. Google Assistant and Amazon Alexa continue to grow in popularity, making voice assistants commonplace in households all across the globe. Even with the momentum of voice rapidly increasing, however, many companies are left wondering where to start.

That’s why Applause brought together some of the leading minds in the industry to discuss the state of voice and provide best practices for developing and testing voice experiences. VoiceTalk San Francisco opened with presentations by Voicebot.ai’s Bret Kinsella and Google’s Cathy Pearl, followed by a fireside chat with Martin Dureja, Product Manager for Voice at Mercedes-Benz. The event closed with a keynote panel featuring Cathy, myself, and Bespoken’s John Kelvie. Throughout the day, attendees learned what it takes to make voice a successful part of their digital strategy.

 

The entire event was filled with best practices and actionable recommendations, but a few lessons stood out that should be kept in mind when starting your voice journey.

Lesson One: Communication is Key

Clearly communicate what your application can and cannot do. Customer frustration comes when users feel they are not being understood. When a customer is greeted with responses like “I don’t understand” or “I didn’t catch that,” they may continue to repeat themselves until frustration boils over and they give up. By providing another response, such as “We do not currently offer that feature,” the customer understands the app’s limits and can move on.

Lesson Two: Do the Heavy Lifting up Front

When developing a new voice experience, putting in extra work up front will lead to an easier implementation process. Developing sample dialogues before you write any code assists in uncovering areas you may have missed. This is also a great opportunity for stakeholders to see a visual representation of your voice experience before development work begins.

Lesson Three: Design for Real Life

Design for the ways that people talk, not the way you wish they did. As Cathy shared, you don’t call it an error when someone says something you didn’t expect or couldn’t understand in real life. Therefore, it stands to reason that you shouldn’t treat challenges in voice conversations as user error. Treat these moments instead as opportunities to improve upon the experience.

Lesson Four: Different Situations Call for Different Experiences

You don’t have to design the exact same experience on each platform. A voice-only experience should have different capabilities than the same app on a smart display. In a smart display experience, think about areas where visuals add to the experience, but don’t try to create a visual for every voice action. Superfluous visuals that do not add additional value can create an inferior user experience compared to having no visual at all. For example, displaying the words someone is speaking on a screen can be a distraction and unnecessary in many instances.

Lesson Five: Use the Natural Accessibility of Voice

One advantage that voice apps have over web and mobile is their natural accessibility. Voice was not originally built for accessibility, nor was it an add-on to an existing product like it is for many mobile and web experiences – it is instead an inherent capability that is uniquely, natively supported by the technology. When companies think about inclusive design, they are primarily thinking about people with permanent cognitive, visual, or hearing problems. However, inclusive design stretches further than that. Every person will face an accessibility challenge at different points throughout their life – whether they are carrying groceries, lost their glasses, or broke their arm. Voice can step in to assist when other applications are impractical or even impossible to use.

VoiceTalk provided a great opportunity for attendees to understand where voice fits in their overarching digital strategy. To learn more about how to build and test great voice experiences, download our recently published whitepaper with Bespoken.

Published On: April 23, 2019
Reading Time: 4 min

How Much Testing Is Enough?

Risk-based testing prioritizes critical tests to reduce risk.

Are AI Tools Improving Accessibility in 2026?

Read the highlights from Applause’s annual survey on the State of Digital Accessibility.

Human Testing vs. AI Testing: What Each Can (and Can’t) Catch

Find the perfect balance for reliable software testing.

From Drift to Deflection: Engineering Trust in AI Systems

Maintaining user trust in your AI chatbots is a continuous process, involving evaluation, observation and adversarial testing.

Test Automation, AI and Gaps in Digital Quality

While AI-generated code and automation can speed releases, they require human oversight to make sure you’re testing what really matters.

What Makes a QA Process Mature?

Mature QA moves from reactive defect-chasing to proactive quality engineering.
No results found.