By this time next year, voice will have moved into the mainstream. Consumers are already adopting voice-enabled devices at a rapid pace. Even a more conservative estimate from BrightEdge expects more than half of consumers to use artificial intelligence (AI) and voice technology in their daily lives by 2020.
While platforms like Amazon Alexa and Google Assistant offer new opportunities, companies need to be careful to build quality experiences. Even the slightest error can lead to frustration or, worse, a negative customer experience. Just think of the phone trees, which really introduced the general public to voice technology through interactive voice responses (and infuriated customers everywhere along the way). Companies can't afford to make similar mistakes with voice apps as early experiences could turn customers off from using voice in the long term.
Today's voice-controlled solutions are making it easier than ever for consumers to engage with companies in every sector, and conversations with AI-enabled systems are continually becoming more expansive and functional.
Several financial services companies are leading the way with voice-enabled chatbots that allow customers to check account balances, make transfers, and send payments, with CapitalOne, U.S. Bank, and ING now offering voice-activated banking solutions through platforms like Alexa and Siri. At the moment, consumer adoption of bill payments via voice is relatively low. However, according to an estimate by Business Insider, the use of voice payments will reach 31 percent of U.S. adults—or 81 million consumers—by 2022.
Voice is set to impact more industries than just banking. A recent study by OC&C Strategy Consultants found that purchases made through devices like Amazon Echo and Google Home are likely to grow from $2 billion in 2018 to more than $40 billion by 2022. When combined with mobile voice-recognition applications, these technologies will become a large extension of customer service.
The growing ubiquity of voice-based devices now offers an opportunity to be omnipresent with customers, all in a form that is immersive and natural for users. Many companies are already creating their own skills for Alexa, but they'll have to innovate for more voice-activated solutions and platforms in the coming years. To build consumer-grade voice experiences, companies need to consider a broad and complex set of variables that include everything from hardware to geolocation, languages and dialects.
A Complex Communication Channel
While voice offers many benefits, it doesn't come without its challenges. Accents and speech patterns can impact the performance of the voice device (and subsequently the customer experience). Unlike text responses that offer clear direction, and AI-enabled chatbots that can easily distinguish the limited character sets and variations of misspelled words, voice is an open sandbox where anything can happen.
With voice, user behavior is also less predictable, and there's little to no screen interaction for richer answers. As a result, consumers don't interact with voice apps the same way they do with GUI apps.
Even the slightest subtlety of a word or inflection of speech can draw the line between a good customer experience and a bad one. And when something does go wrong, it can be difficult for companies to determine if subpar performance is the result of a bug or a user request that is not understood.
Testing such complexities remains one of the biggest bottlenecks and barriers to building successful voice apps. Even the largest companies are still refining their offerings, which is why many companies have little used or poorly performing skills on the market. To that end, according to a recent article in VentureBeat, 60 percent of Alexa's 30,000 skills have no ratings at all. Testing is the key to ensuring voice works for all customers, on all devices, in all locations. The most important thing is to test the user experience during prototyping to make sure that users understand the concept, find the navigation and experience intuitive, and can complete expected tasks.
Three top methods for testing the entire voice experience are dialog verification, functional testing, and user experience validation.
Dialog verification ensures commands are understood by voice apps by leveraging real users across a breadth of languages and dialects to validate the speech recognition skills of apps and enabled devices.
Functional testing validates that every possible flow through a voice experience on different combinations of devices and operating systems work as intended in real-world scenarios.
User experience validation gathers subjective feedback from actual users to identify trends that might frustrate customers and cause them to stop using voice.>
However they innovate and test, companies must develop a strong strategy for voice that is driven by the goals and persona of the organization. With so much at stake and so many complexities in the channel, it is essential to leverage a variety of testing types to ensure the high performance of voice quality from the start.
Chris Sheehan is vice president of enterprise customer success at Applause, responsible for developing and executing the company's retention and expansion programs for North American enterprise customers. Prior to his role in customer success, Chris led Applause's long-term Product Strategy team. He has also led many investments in the Boston area as a software venture capitalist.