Every year, Localytics hosts ENGAGE, a conference for mobile marketing leaders to come together and talk mobile marketing strategy and innovation. This year, Rainer Leeb, Senior Director of Growth, took us through how voice recognition is driving A.I. at Localytics customer SoundHound.
Today? Rainer Leeb of Soundhound says that you need a voice-enabled A.I. strategy–or fall behind.
The rise of the digital assistant means that more than ever, we’re accessing apps through Alexa, or Siri, or Cortana, or Google Home. And now, there’s a new digital assistant that’s building buzz: SoundHound’s native assistant Hound.
SoundHound’s been working on their voice technology for over 9 years, and pioneered a new approach to speech recognition. Traditionally, Automated Speech Recognition (ASR) and Natural Language Processing (NLU) are two separate processes speech recognition software has to do in order to understand human speech. Hound combines these two processes into one step, making the platform faster.
There’s also the question of usability and machine learning. If you’re a frequent Alexa or Siri user, you know how you construct your question is pretty critical. Say you’re looking for a place to grab dinner. You know you don’t want pizza, and you want to make sure you’re only looking at places that are currently open. Most virtual assistants can’t exclude categories, or drill down further into a topic. But Hound can. It’s this contextual accuracy which puts it leaps ahead of other assistants on the market.
But, as Leeb notes, the experience isn’t seamless (yet). It will take time for users to learn how to effectively interact with artificial intelligence platforms like Hound, and for these platforms to learn enough from these interactions to bridge the gaps–just like search engines.
Rainer Leeb, Senior Director of Growth at SoundHound, and our trusty ENGAGE team: Senior Director of Marketing Communications Kristin Cronin and Senior Marketing Manager Susie Castillo.
How should a virtual assistant platform work?
The right platform, Leeb told us, is independent and agnostic. It doesn’t compete with the integration using it. Users can maintain their own brand, and create customized experiences for their own users. Voice-recognition platform themselves will win with more speed, accuracy, and understanding of complex domains.
Thanks to Rainer for a fascinating session! Try out Hound for yourself on SoundHound’s app, and see the full list of ENGAGE 2017 sessions here.