John Kelvie is the former co-founding CTO for both Overture software and XAPPmedia. In the latter, John actually built a speech recognition engine for use in mobile apps. Those experiences taught him a lot about supporting resilient enterprise scale systems and voice interaction respectively. They also forged his insight around key gaps for voice application development, management and support and led directly to the founding of Bespoken. Bespoken is a software company that builds open source voice application development tools and provides a SaaS solution for logging, monitoring, alerting and validating voice applications in production.
In this week’s podcast, John talks about the architecture of voice systems, the impact AI platforms have on developers and what is missing from the current voice technology stack. He also weighs in on the difference between speech recognition for wake words compared to intent utterances. And, of course, we talk about monitoring voice applications and the future of voice assistants. You are sure to learn a number of things about voice technology from the interview.
- Find John Kelvie on twitter here: @jpkbst
- You can learn more about Bespoken here: https://bespoken.io
You can listen to the podcast interview above or at any of the locations below.
This interview was conducted in the second quarter of 2017, but all of the content is still relevant and on point. One thing that has changed is Amazon this past Friday offered a new approach to develop Alexa skills for children that comply with COPPA. More on that tomorrow.
Also, we mention in the discussion an earlier Voicebot interview with Mike Elgan. Well worth a read about the future of voice and computing.