AiThority Interview With Scott Stephenson, Co-Founder and CEO at Deepgram
Hi, Scott. Please tell us about your journey into technology and how you started at Deepgram.
The idea for Deepgram was born two miles underground while I was a Ph.D. student researching the creation of dark matter. In the hours not devoted to research, my co-founder Noah Shutty, and I recorded audio from our lives, 24/7. Yet, when we tried to go back and find key conversations in those audio files, we were left with countless hours of silence and background noise that was almost impossible to sift through. We quickly realized that there wasn’t a tool that would allow us to process our recordings and pinpoint valuable timestamps. Noah and I used AI techniques to find key moments buried within large datasets of recorded audio – this was the very beginning of Deepgram.
What is Deepgram? What does your Ideal Customer Profile look like?
Deepgram is an Automatic Speech Recognition (ASR) platform, powered by AI. We’ve rebuilt the entire speech processing stack, ditching traditional data processing pipelines, Hidden Markov models and heuristics for end-to-end deep learning. Our Deep Neural Network (DNN) utilizes Convolutional (CNN) and Recurrent Neural Networks (RNN) to deliver an affordable, highly accurate, and insanely scalable speech solution. Typical customers are large enterprises looking to replace manual efforts in call centers, and software companies (emerging and established businesses) that want to incorporate speech into their product offering.
Tell us more about Deepgram’s AI and NLP capabilities. How can marketing and customer service teams derive maximum benefit from these two techniques?
Deepgram delivers enterprise-grade speech recognition and understanding at scale. We do it by providing patented model training and innovative data-labeling alongside flexible API deployment options. With our platform companies can train a model to understand their unique audio, and with ongoing training sessions improve accuracy.
To accurately classify different audio types, you need an accurate transcription. Additional NLP features that we offer on top of ASR include Multi-Channel support and Speaker Diarization.
How do NLP- driven companies like Deepgram address customer pain points in the COVID-19 era?
Prior to the pandemic, consumers had already widely adopted voice technology with the use of smart speakers, smartwatches, etc. While voice technology began as a nice to have, it has now become a necessity for businesses to survive. In the same way that an omnichannel strategy is now essential for new companies to survive especially in the retail, food and baking spaces, voice will become a competitive advantage among enterprises. Voice technology is a more intuitive way to communicate for consumers because they are using their unique voice to describe what they need vs. interacting with a chatbot, which can often make the customer experience feel disconnected.
IVR is changing the call center ecosystem. What’s the future of Speech recognition platforms in the remote collaboration era?
IVR is an essential ingredient of all modern contact centers. The benefit of IVR is that customers are able to use their voice to navigate the menu versus using a touch-tone, which requires customers to listen to an automated menu and press multiple numbers before being routed to the correct customer service agent. It’s great in theory, but not all customer inquiries fit into a multiple-choice menu. The experience can easily turn from helpful to frustrating when the system struggles to accurately capture the reason customers are calling or the customer has to listen closely for the correct menu prompts instead of being able to simply state the reason for their call.
Integrating Deepgram into existing IVR solutions minimizes frustration for first-time callers by ensuring that they are heard correctly the first time. Deepgram provides our customers with unparalleled accuracy (over 90%), taking into account keywords that are important to your customers and automatically adjusting to noise pollution, meaning that the customer is heard the first time, regardless of background noise.
Tell us more about your remote workplace technology stack? Which tools/platforms are you currently using for Marketing, Sales and Communications?
Deepgram relies heavily on solutions like Slack (also an investor of Deepgram), Zoom, Google Suite, Gong.io, Salesforce, HubSpot, Asana and Linear on the development side. We are also looking into employee engagement solutions as Google Drive is not easy to navigate for new hires and sharing information company-wide.
What is the most challenging mission you have undertaken in the last 6-8 months / or, you plan to take in the next six months?
Deepgram received our Series A in March and has been scaling up our sales, marketing, customer success, research and engineering functions. Ensuring every employee has a seamless onboarding experience and can quickly become productive is our #1 priority. As more employees come on board, new products are also being released, and new customers are joining the platform. We are experiencing growth from every aspect of our business and will be making sure processes are improved internally so we can better serve our customers.
Tell us about an outstanding digital marketing campaign at Deepgram. How did you plan the project with your CMO and measure its ROI/ effectiveness?
Katie Byrne, our VP of Marketing joined us in May and within three months launched a new website. We needed a value proposition and supporting design that aligned with our developer and Enterprise target audiences. The website also needed to lay the foundation for lead generation both through organic SEO and paid advertising. We broke the website launch into two phases to align with two product announcements (AutoML, UniMRCP). This ensured we had our best foot forward. Already we have seen an increase in inbound lead flow and are landing opportunities in our top Enterprise and Software accounts.
Hear it from the pro: 3 key factors that drive a positive customer experience overall:
Three key factors that drive a positive customer experience are timeliness, convenience and technology. Whether in person or on the phone customer inquiries should be addressed in a timely manner. No one likes waiting in line to speak with a representative or being placed on hold – having a timely response time will support a positive customer experience.
Getting in contact with a customer service representative should also be convenient. There shouldn’t be multiple menus for customers to listen and respond to in order to be routed to the proper customer service agent – implementing automatic speech recognition into your customer service strategy will ensure that your customers are heard correctly the first time.
It’s also important to conduct an audit of your current customer service hardware and software and recognize where there are areas to improve the overall customer experience through technology. Deepgram provides unparalleled accuracy (over 90%) regardless of background noise and takes into account the unique characteristics of an individual’s voice ensuring that customers are heard correctly the first time.
Any advice to every young marketing and sales professionals in the tech industry?
Working at a technology company is hard – especially a startup. There will be ups and downs and constant changes, but if you are connected to why a company exists, that passion will help you sail through the stormy periods.
Tag a person in the industry whose answers you would like to see here:
Thank you, Scott! That was fun and we hope to see you back on AiThority.com soon.
Scott Stephenson’s bio: Scott Stephenson is a dark matter physicist turned Deep Learning entrepreneur. He earned a Ph.D. in particle physics from the University of Michigan where his research involved building a lab two miles underground to detect dark matter. Scott left his physics post-doc research position to found Deepgram.
Deepgram is the leader in enterprise automatic speech recognition (ASR) for call centers and software providers. With our patented end-to-end deep learning approach, data scientists get access to the industry’s fastest, most accurate and highly scalable AI technology. We take the heavy lifting out of noisy, multi-speaker, hard to understand audio transcription, so you can focus on what you do best.