The technology undergoes a transformation as omnichannel takes over interactions
The field of voice interaction design is continuing to evolve so that even people with very limited technical prowess or coding knowledge can now quickly and easily create systems to handle very basic speech interactions or work with other systems in an omnichannel environment. The common goal among vendors of these technologies is to help companies provide their customers with rich, voice-enabled experiences that aren’t overly complicated, costly, or time-consuming to create, deploy, or maintain.
Much of the current work has centered on expanding beyond very basic capabilities, such as the voice inputs given to interactive voice response systems to retrieve account balances or flight times, to more advanced solutions that involve other aspects of the contact center, leading to better customer experiences that cross channels.
There are two separate approaches: One looks at the sheer voice dialogue, and the other looks at more comprehensive interface design, says Deborah Dahl, principal of Conversational Technologies, a member of the VoiceXML Forum, and chair of the World Wide Web Consortium’s Multimodal Interactions Working Group.
Among the tools in play is Microsoft’s Visio, a diagramming and vector graphics application that is part of the Microsoft Office family and still widely used for a number of advanced voice interactions, according to several experts. But the Visio platform, which was first launched in 1992 by Shapeware, a company Microsoft later acquired, was never designed for voice, Dahl adds.
Another option for companies looking to deploy voice was to use the design tools that came with their contact center platforms—in a way skipping the design phase entirely, explains Kerry Robinson, CEO of VoxGen, a London-based provider of IVR solutions and services. Alternatively, some people used a combination of Word documents and Visio.
“The problem that we see with this, and what we hear from our customers and our prospects, is that though Word and Visio are very flexible tools, they weren’t designed for managing the design of complex voice applications. As a result, [customers and prospects] often find that their documentation gets out of synch with what is actually deployed,” Robinson says.
For years, the speech community has tried in vain to look past Visio. Today, that’s not such a pie-in-the-sky ideal anymore. Real alternatives have finally started to emerge in the past few years.
One alternative has been the Voice User Interface Design (VUID) Toolkit, an open-source graphical platform for text or voice user interface design developed in 2007 by Matt Shomphe, lead speech scientist at Convergys, and Phil Shinn, now a senior user interface designer at VoxGen.
Like Voice XML, a standard programming language for developing audio and voice response applications first released in 2000, the VUID Toolkit was designed with voice in mind and offers superior capabilities.
“The newest version [of the VUID Toolkit] allows for real-time collaboration and integrates with the testing platform,” Dahl says.
The testing available through the VUID Toolkit is automated, making it smoother and less labor-intensive than older testing methods, according to Dahl. Testing, she says, is an essential part of any project. “You have to keep testing as you add new features during the design to make sure that you haven’t lost any functionality.”
Real-time collaboration is another important advantage of using the VUID Toolkit, Dahl adds. “This is a natural part of the growth of the underlying tools. After the launch of Google Docs, remote collaboration is moving into the industry like it has everywhere else.”
The increased opportunity for collaboration has given rise to a number of industry groups, including the Voice User Designers Association and the Association for Voice Interaction Design (AVIxD), which in 2016 defined issues and questions around conversational user interfaces. AVIxD’s next step will be to work on design guidelines for those interfaces.
As it does that, conversational speech and natural language understanding will continue to be important to the advancement of voice design, Dahl points out.
Another contender to come on strong in the past few years has been VoxGen, which has upgraded its user interface design tools to include the following:
• Documenting, to log design specifications and problems with transitions, grammars, etc., that might come up along the way.
• Prototyping, allowing companies to test designs with end users before putting anything into development. Companies can invite end users to try the system over the phone while a designer behind the scenes presses the appropriate buttons on the prototype.