Scott Stephenson, CEO of Deepgram, discusses the evolution and business applications of Voice AI, exploring its value in applications and common use cases. He also delves into the process of adding Voice AI to existing applications and collecting user feedback for voice-centric interactions.
36:18
forum Ask episode
web_stories AI Snips
view_agenda Chapters
auto_awesome Transcript
info_circle Episode notes
question_answer ANECDOTE
From Dark Matter To Audio AI
Scott Stephenson moved from building deep underground dark matter detectors to founding DeepGram after realizing audio signal processing skills transferred to speech AI.
He and his team recorded and analyzed waveforms continuously, which sparked the idea to index and search audio at scale.
insights INSIGHT
Voice AI Has Turned On
Voice AI has reached a turning point where quality, latency, and cognitive ability combine to make interactions feel real.
Perception (speech-to-text) is mature, and rapid improvements in TTS and LLMs now enable human-like, low-latency voice experiences.
insights INSIGHT
Cost Determines Real-World Viability
Cost parity with human labor is essential: voice AI must stay below approximate human-hour costs to be competitive.
Achieving low enough inference and orchestration cost is as important as technical quality for adoption.
Get the Snipd Podcast app to discover more snips from this episode
Topic 1 - Welcome to the show. Before we jump into Deepgram and Voice AI, let’s talk a little bit about your background and then what led you to start the company.
Topic 2 - Let’s start by stepping back and looking at the bigger picture. Many of us have interacted with voice technologies, from Siri to Alexa to various call-center “agents”. From a maturity perspective, where are we in terms of voice-centric interactions with computing systems?
Topic 3 - When businesses think about including computer-aided voice-centric interactions to their applications, what’s the typical thought process for where it makes sense to the business, how it represents their brand, etc.?
Topic 4 - How does Deepgram bring Voice AI to applications? What do developers have to think about when adding Voice AI to an existing application, or to think about building a voice-centric application from the start?
Topic 5 - What are some of the most common use-cases when Voice AI adds new value to the application? What areas are starting to push the edges of possible?
Topic 6 - How do companies usually go about collecting user-feedback about the voice-centric interactions?
Topic 7 - If you look forward a year or two, where do you see Voice AI types of technologies evolving?
The Cloudcast