The State of Speech Developer Platforms
Consequently, uses for speech engines are growing. Financial services companies are turning to speech to make it easier for customers to use their services. American Express has teamed up with Alexa so consumers can use their voice to make an online payment. Capital One created a skill that tells customers how much they spent in a time period or on an item. Travelers Insurance’s speech functions enable new Auto ID cards to be mailed to clients. Credit Card Helper finds the best credit cards for cash back, rewards, travel, and building credit.
Applications are emerging in other areas. The Voice of Social Sales application features Chelsea Peitz, an author and real estate industry veteran who offers lead generation advice. Comcast Business ActiveCore provides customers with an understanding of how well their enterprise network is functioning and outlines how features like dynamic routing impact performance.
A Look Ahead
As the ecosystem expands, management becomes a more vexing challenge. The first step is building management functions in speech applications. Some steps have been taken; for instance, Dialogflow’s tool links to Google’s Stackdriver monitoring, logging, and diagnostics solutions.
But as speech applications become more dispersed, the underlying infrastructure grows much more complicated. Many of the new solutions are cloud-based, which creates two challenges.
Management is the first area. “Currently, vendors do not offer much analytic help if an application encountered problems such as latency,” explains Dahl. Enterprises need tools to help them understand what impact speech applications have on their infrastructure. If a speech skill is getting bogged down, it often becomes difficult for the company to trouble-shoot and fix the problem.
Second, the cloud is becoming a common foundation for new speech features. But most firms have a lot of time, money, and expertise expended on legacy systems. Consequently, these new capabilities need to be tied into the existing data center infrastructure. “Customers want speech applications that use legacy system information as well as cloud capabilities,” says Roy Lindemann, cofounder and CMO at ReadSpeaker, a subsidiary of Hoya.
Currently, the industry lacks clear standards, so the connections are complex and rare and only become available in an ad hoc manner—when a vendor or third party builds them. Corporations with multiple solutions expend many resources gathering management data and making changes when those systems are updated. Streamlining such processes is on their wish list for 2019. Hoya has taken on the management challenge itself. In the past two years, the company acquired Neospeech, ReadSpeaker, and Voiceware, which supply speech recognition solutions. “To enable customers to connect our system to other solutions, we need to first provide consistency among our own interfaces,” noted ReadSpeaker’s Lindemann.
The speech development market has rapidly evolved. Vendors have put many building blocks in place, but new requirements have emerged. As businesses deploy more speech applications, they are looking for help in areas like integrating and management.
Paul Korzeniowski is a freelance writer who specializes in technology issues. He has been covering speech technology issues for more than two decades, is based in Sudbury, Mass., and can be reached at firstname.lastname@example.org or on Twitter @PaulKorzeniowski.
Want to understand what is happening with speech application development platforms? Then take a close look at smartphone application stores.
As we've done now for the past few years, Speech Technology magazine is again dedicating its first issue of the new year to a preview of what's to come in our small corner of the world. We've highlighted the six technology areas where we see the most impact: speech engine, speech analytics, voice biometrics, virtual assistants, speech developer platforms, and assistive technologies.
Finding a Voice in the IoT