This is a fantastic discussion on how to cold start an AI company, build it to scale, etc. Also, excellent guidance on white spaces around which to build new AI companies.
Sramana Mitra: Let’s start by introducing our audience to yourself. Tell us a bit about your background and all also introduce us to People.ai.
Oleg Rogynskyy: I am the CEO and Founder of People.ai. I have been doing startups all my life. My previous company Semantria was also an AI company. I started it in 2011 and then sold it in 2014. People.ai was started in 2016 when I moved out here to Silicon Valley. I have been working on it ever since.
Sramana Mitra: Where did you move from?
Oleg Rogynskyy: Montreal. I am originally from Slovenia. I worked for one of the first AI companies called Nstein Technologies from 2006 to 2010. We sold it to Open Text. Our technology now is the Open Text AI. I left that to start Semantria in 2011 and the rest is history.
Sramana Mitra: What was the premise of Semantria?
Oleg Rogynskyy: Semantria was the world’s first sentiment analysis cloud API. You could send any piece of text, a tweet, survey, or document, and within a second, it would come back to you to show the sentiment in this text as well as the people, places, companies, and categories of all the metadata you would need to build a product on top of that tweet.
Sramana Mitra: Who did you sell it to?
Oleg Rogynskyy: We sold it to a company called Lexalytics.
Sramana Mitra: We’ve covered Lexalytics. I am familiar with them.
Oleg Rogynskyy: They were the big players in the space, and they wanted to get into the cloud game because they were on-premise at that time. We were the cloud solution on which they are still doing a significant portion of their business.
Sramana Mitra: Let’s come to People.ai. When did you start People.ai and what was your observation at the beginning?
Oleg Rogynskyy: People.ai was started in February of 2016. We just turned five years old. It felt like one day. The initial vision of why I wanted to start this company came through my employment history. I will never forget this day. Back in Montreal, in 2007, I was in Nstein and our Chief Operating Officer grounded me and the whole sales team for a week to clean Salesforce. We literally spent a week in a conference without windows, googling companies and people to update them on our Salesforce.
Cleaning Salesforce lasted for about three weeks and then after that, it was back to being a mess. That was the first learning of saying, “Hey, there is a problem here.” During Semantria days, one of the business lines that we were doing was e-discovery. We were using our AI and draft engine to understand massive amounts of email at scale to help find the bad guys.
That was when I learned the power of natural language processing as it applied to email and calendar data at a large scale. When I sold Semantria, I worked for a company that you might be familiar with. It’s called H20.ai.
Sramana Mitra: Yes.
Oleg Rogynskyy: I was running sales and marketing for H20 for a bit. H20 products are spectacular from the designer’s perspective. That is where I learned all the modern techniques of doing machine learning at a significant scale. Combine that with the understanding that there is a big problem with Salesforce data and how companies are flying blind as well as the understanding that there are a lot of interesting nuggets and insights hidden in every company’s email, calendar, Zoom activity data. With modern technology from H20, I knew that if I put it all together, there would be a big business here.