The Voicebot Podcast

Harjinder Sandhu founded Saykara in 2015 but his journey into healthcare and voice began in 2000 when he founded Medremote which used speech recognition and machine learning to automate physician notes transcription for medical documentation. That company was eventually acquired by Nuance in 2006 and Sandhu stayed on as the healthcare group CTO until 2011. He then was a co-founder of a patient engagement platform called Twistle which is still operating today. But, the draw of voice technology advances led him to Saykara where he and his team are building a voice assistant for doctors and other clinicians. It's a great conversation about specialty assistants, the importance of workflow in healthcare, and the tech stack required to serve this highly demanding user base. Sandhu earned a PhD in Computer Science from the University of Toronto and was a professor of computer science at York University before becoming a startup founder.

Todd Mozer is founder, Chairman, and CEO of Sensory. The company has provided voice and AI technology for 3 billion devices over 25 years. Sensory's technology stack can be used to deploy a custom assistant in an appliance or mobile app and also has individual features for wake word, biometric login, and more. Sensory customers include Amazon, AT&T, Garmin, Google, GoPro, Huawei, Motorola, and Sony to name a few. Todd is a graduate of Stanford Business School and UC Santa Barbara. He also studied computer science at UC Berkeley in between.

Peter Durlach started out in Audiology technology in the 1980s and then quickly transitioned to speech technology in 1988 at Articulate Systems which was acquired by Lernout and Hauspie in 1999. He then moved on to lead Unveil Technologies, a pioneer in the contact center segment. Eventually, Durlach wound up at Nuance twice with a brief interlude at the hospital system UPMC. Today, Durlach leads Nuance's $1 billion healthcare business which includes solutions used by over 85% of radiologists and 65% of physicians that incorporate speech recognition, dictation, and custom assistants into their daily practice activities. In our conversation, we discuss Pete's three decades in the speech and voice technology, building custom assistants, and where voice is and will be making a big difference in healthcare.

Scott Stephenson is cofounder and CEO of Deepgram. He earned a PhD in particle physics at the University of Michigan with a focus on Machine Learning. He then spent time building a system designed to detect dark matter two miles underground. That experience inadvertently introduced Scott and a colleague to the limitations of speech recognition systems which to their surprise were not using the same deep learning architecture with hardware acceleration they had built to detect dark matter activity. It ultimately led to founding Deepgram in 2015 as a new model for automated speech recognition. The company added 12 million in funding in March of 2020.

Alan Nichol co-founded Rasa in December 2016. Rasa began life as an open source NLU for chatbots and has expanded to other features, including voice interactivity. The company raised $13 million in venture funding in 2019 and came back about twelve months later with a $26 million funding round in 2020 after showing strong growth numbers in both open source downloads and new enterprise clients. Nichol is CTO at Rasa and before that was co-founder and CTO of a Techstars-backed productivity startup that had a conversational angle. He earned a PhD in Engineering from Cambridge with a machine learning focus. He also has an advanced degree in chemical physics from the University of Edinburgh. Today, we talk about the conversational AI stack, chatbots, voice, open source, and where we are headed with custom assistants.

Punit Soni is founder and CEO of Suki. The company provides a specialty voice assistant for doctors and other clinicians and he elaborates how its features go beyond the earlier technologies associated with medical dictation and transcription. The company was founded in 2017 and has raised more than $40 million since that time. Earlier, Punit was Chief Product Officer at Flipkart and Vice President of Product Management at Motorola while it was a division of Google. At Motorola, he oversaw the rollout of Moto X which was one of the first smartphones to include a voice assistant. He was also a Lead Product Manager at Google and began his career as a software engineer. Soni earned an MBA from Wharton and an MS in Electrical Engineering from the University of Wyoming.

Mike Zagorsek is vice president of product marketing at SoundHound. Prior to SoundHound, Zagorsek was head of product marketing at Square and Leap Motion, and before that was Director of Interactive at Apple. Way back in 1999, he was a Manager at Sapient where he helped companies enter the digital age much like he is now ushering them into the voice age. SoundHound is a leading white label solution for building custom voice assistants and has made news recently for powering Snapchat's new assistant. In today's interview, we discuss that implementation along with Pandora, Mercedes, and others.

Voice at Swisscom began in 2015 with a voice interactive TV remote. The voice assistant Hey Swisscom emerged five years later in November 2019. Misha Zivkovic and Riccardo Lopetrone learned a lot in both versions of their voice-enabled products and they share the details behind the voice assistant in today's extended discussion. Do you think you could design and build a custom assistant in 9 months? Swisscom just did. We discuss their journey with voice, tech stack, objectives, and more. Misha Zivkovic is Program Manager for voice at Swisscom and Riccardo Lopetrone is Senior Product Manager for Voice, a role he began in 2015.

Andy Webb is head of product for voice and AI at BBC Worldwide as VP of Product Strategy. BBC has been a pioneer in supporting consumer assistants globally such as Amazon's Alexa and Google Assistant. It is prominent in the UK but also provides skills and Actions in many other countries as well. It is also one of the most respected news organizations worldwide so Amazon and Google are incentivized to make sure BBC is available through there assistants. Last August, BBC announced it would create a new voice assistant of its own called Beeb. The organization is not abandoning Alexa or Google Assistant. Instead, Webb tells me that they are supplementing their work on these platforms. Beeb is more than an assistant. It's BBC's own internal conversational engine that can enable custom assistants deployed through desktop and mobile apps as well as support the many voice assistants that want access to BBC content. Media is one of the most popular and most important segments for voice assistants today. BBC is developing the Beeb to offer richer functionality than you could ever provide on the general-purpose consumer assistant platforms because of the limitations they impose on third-parties. Andy Webb joined BBC in 2106 and has been head of the voice and AI group since it was formed in 2017. Before his time at BBC, Webb was EMEA regional manager for Imagine Communications, head of product management for digital platforms at Arqiva and Product Development lead at Sky among other roles.

My guest today is Ramu Sunkara, co-founder and CEO of Alan AI. In 2014, Ramu co-founded a company called Synqq which was a voice assistant for meetings. Sometimes startups are for learning. Synqq educated Ramu and his co-founder that enterprises truly needed voice for enterprise apps. That led to the creation of Alan AI which has been his focus since 2017. Earlier in his career, Ramu was co-founder and CEO of Qik which was a video streaming solution for mobile and was acquired by Skype in 2011. Before that, he spent 11 years at Oracle, serving in, among other roles, vice president of e-Business suite technology, and vice president of real-time collaboration technologies. Ramu was trained as a software engineer and holds a master's degree in computer science from the University of Wisconsin and earned an MBA at Boston University.