The conference

At the end of April, I went to Washington DC to attend SpeechTEK 2019, the world’s leading speech technology and AI conference. The event was exactly how I’d imagined with tracks covering all aspects of speech technology with leading experts giving keynote talks on delivering and deploying speech technologies in all types of applications.

My objectives from attending SpeechTEK.

To learn more about what people are doing with speech, to determine where it adds value and to find some other cool applications or use cases of the technology. I also wanted to meet some interesting people in the industry and the event certainly didn’t disappoint.

What did I learn?

I attended some really great keynote talks presented by leading industry experts. I have outlined notable takeaways below.

Algorithms In, Humans out – Jarno Duursma

Takeaways

• There are different ways of humanising AI: language, vision, voice and emotion

• Trends in AI are largely focused around assistance and autonomous AI

• Some drawbacks of AI include software for generating fake artifacts

• Education and creativity are ways of dealing with the drawbacks

Will AI Replace Humans in Customer Engagement? – Andy Peart

Takeaways

• It is unlikely, in fact the number of jobs are likely to increase

• It is hard to fill in AI jobs

• AI is likely to help humans do their jobs better

Speech in the Car – Thomas Schalk

Takeaways

• Speech recognition is the number one thing drivers complain about

• There is limited footprint to run ASR in the car

• Car manufacturers are opposed to people using mobiles in the car and want drivers to use systems in the car directly

There was also an interesting discussion on the contact centre landscape. As one of Speechmatics’ target markets, it was interesting to hear an expert’s perspective that the aim of contact centre software is to improve the experience of both the agent and the customer. We couldn’t agree more with this statement having partnered with a number of contact centres to drive agent empowerment and improve customer experience.

Overall, I learned that ASR isn’t a thing in and of itself. To generate true value, ASR must be integrated into some larger system. Often when people think of speech they think of virtual assistants, dialogue systems or speech analytics, but in reality speech technology is part of a much larger ecosystem that is continually growing.

Another talk that grabbed my attention was one on the use of speech recognition in space. I have to admit, it isn’t a use case that I have given much thought to, but it was interesting to hear from NASA about how speech can assist astronauts: voice control of the shuttle CCTV system, EVA retriever and EVA suit control, for example.

Speech Tech Magazine Awards

On a slight side note from the conference itself, before heading out to Washington DC, Speechmatics were announced winners for the Best Speech Recognition category in the 2019 Speech Tech Magazine's People's Choice Awards. The award win is fantastic recognition of what the team is doing here, helping to shape the future of voice for any application. Who knows, perhaps one day our technology will be used by NASA and for those astronauts in deep space.

Interested to learn more about how speech recognition technology can work for your business? Get in touch!

David Mrva, Speechmatics