Ok, Google: Tell me about Israel’s voice recognition technology
Startup Nation is one of the biggest hubs for voice recognition technology - and different solutions make the possibilities endless
Siri or Alexa might help us with our productivity tasks, but the technology goes back far earlier than 2011, when Siri was born, and the possibilities reach much further. CTech spoke with Galit Galperin, a product executive at Playtech, a lecturer at IDC Herzliya, and the author of the popular industry blog Voicey, about where the technology came from, how Israel’s startup community creates unique solutions, and how it “became the love of my life.”
“This is not a new technology,” she explains when talking about virtual assistants. “20 years ago we had IVR (interactive voice response) systems that were using voice recognition in a small sense. You’d say your name, or whatever number you needed to say, and it would recognize you.”
Today, it has become far more advanced and more likely to mirror the type of technology we were used to seeing in Sci-Fi films and shows as children. “We have a few different layers: it’s voice recognition, it’s natural language understanding, it’s the design of the conversation itself, it’s the whole thing. Nowadays you can have a sort of communication which goes both ways with the machine.” The industry collectively refers to it as Conversational AI.
Many of us are most familiar with the ‘call and response’ type of virtual assistants on a consumer level through our phones or smart speakers. According to Galperin, our dependence on these tools has risen since the coronavirus (Covid-19) pandemic sent us all away to work from our homes. However, she listed a few different companies that are currently making a splash on the commercial B2B scene and whose applications go far beyond the Turing test:
Audioburst encourages ‘next-gen’ listening by listening to millions of minutes from podcasts and radio stations each month to create short audio clips using artificial intelligence. Their Natural Language Processing (NLP) technology seeks specific keywords or speakers to help deliver digestible snippets, or ‘bursts’, used to help with audio editing and distribution.
Galperin: “As we are seeing in podcasts, this is something that is becoming very popular. They have potential.”
Tukuoro operates ‘at the edge’ to bring voice activation to any device in public areas and indoor spaces. Its Tukuoro Voice Open Platform API can be integrated into kiosks, ticket stands, or healthcare devices. Most importantly, it operates outside of ‘The Four’ and values the privacy of its users by not sharing data.
Galperin: “This is fascinating for me, in terms of the potential. Especially in Covid-19 days, when you want to order something and you don’t want to touch the screen.”
We’ve all had conversations with machines, and we know this because it’s so obvious when we talk to them. Voca.ai automates and scales customer conversations with ‘friendly voice AI agents’ - meaning, you’re likely talking to an intelligence system next time you are on a call with a call center implementing Voca.ai’s solution.
Galperin: “Google just signed with Verizon on things like this. It has potential in terms of natural language understanding.”
Voice is more consumer-facing and helps those with severe speech impediments speak via a mobile or tablet app. Using artificial intelligence, the system learns how each user speaks and then builds up a vocabulary to help translate slurred or unintelligible speech patterns. In July 2020, the company raised $10 million in series A.
Galperin: “If you can take someone with a speech disability and understand and translate it, this is an amazing solution and tool. This is like using voice recognition to do good in the world.”
Verbit is a smart transcription and captioning solution for healthcare professionals, universities, and journalists. Its speech-to-text technology is 99% effective and can be invaluable for organizations that rely on transcription at a fast and effective pace.
Galperin: “Imagine you are in a doctor’s office and instead of waiting for the doctor to write everything, the machine listens and transcribes everything for the doctor. This is amazing.”
Gong.IO helps sales teams prepare for their calls with potential customers by recording phone conversations and analyzing the way each person speaks. It then provides feedback on what language and tone to use for specific marketing personas.
Galperin: “The machine records me, and analyzes, and gives me the proper way to speak with the right personas in order to make the sale.”
After outlining all the potential companies coming out of Israel, it became clear that the latest instrument in the tech sphere will not be a device or a software download, but our own organic tool that requires no guidebook. “You had touchscreen technology, and you needed to learn how to use swipe and how to zoom. With voice: it’s something you know from the moment you’re born. You are hearing your mother’s voice, you know how to speak the language.”
There are many use cases that show that voice technology is not just here to stay, but it’s about to thrive. Now Siri, write that down.