Whitney Durmick: Igor, you are no stranger to the entrepreneurial space. Do you want to share a little bit of your startup background?
Igor Jablokov: Sure, I actually started my career as a research engineer and then departed when IBM was not being as aggressive as I wanted them to be to go into the AI space with early Watson. And so I created a venture-backed company and it ended up becoming Amazon’s first-ever AI-related acquisition, which many people now know as Alexa.
WD: That’s an incredible bona fide to have under your belt. But let’s talk Pryon: Pryon is an AI-based platform that helps companies specifically with knowledge management, when did you recognize that this was something that was needed?
IJ: While all of us had these great AI assistants in our homes and our automobiles and on mobile, I decided to catch our on football in that it would be really nice to have that same natural experience for running more enterprise-style workloads, but it wasn’t going to come in the same way as a device sitting in our kitchens or by our beds. Enterprises are going to have much more robust requirements for accuracy, scale, security and latency. And so when we started looking at that, we noticed there aren’t AI systems that are prevalent.
What’s one problem that is common to all of these enterprises across industries that they haven’t solved yet? Knowledge management. And the reason for that is, the content that customers need is strewn about everywhere. They’re in SharePoint, they’re in s3 buckets. They’re in applications like Oracle Financials, SAP. They’re in SharePoint, in Confluence. And so there wasn’t one source of truth.
We saw that being an opportunity to essentially harmonize all of this under a single natural language interface.
WD: So for instance, if I’m a customer service rep, I’m probably getting a lot of the same questions over and over. Would Pryon help me find easier answers and optimize that process?
IJ: Absolutely. Instead of doing an enterprise search, where you type in some keywords and you get a list of documents and you are probably putting your client on hold, for those 2 minutes, you’re in terror, trying to speed-read what the possible solution might be.
Think of the AI then being augmented intelligence, helping you through that journey. When they first come in through the IVR, maybe it’s the Genesis IVR, or something of that sort, it could attempt to answer the question directly, or a chatbot can attempt to answer the question directly before it ever gets to you. But if it does open up a ticket, the minute you open up a ticket, the AI can attempt to automatically answer it. But then if it finally gets to you, and it’s a more complex problem, you can actually be querying the AI while you’re working with the client. And it can be giving you possible solutions quickly and rapidly while you’re engaging with the phone so nobody has to be on hold. And this just increases customer sat and it really reduces your workload and stress level.
“So many people think AI is like science fiction, ‘Oh my gosh, AI is coming, all of us are going to be out of work….’ I think that’s hogwash. I think humans are still the center of creativity, of innovation and intuition. I prefer to think of it as a partnership with machines, not as anything that displaces people.”
WD: So do you agree that there needs to be a bit of a balance between the AI and the human side, there’s really only so much that an AI can solve for; ultimately, there needs to be that human connection when we’re solving problems for customers. Do you agree with that?
IJ: I completely agree. So many people think AI is like science fiction, ‘Oh my gosh, AI is coming, all of us are going to be out of work….’ I think that’s hogwash. I think humans are still the center of creativity, of innovation and intuition. I prefer to think of it as a partnership with machines, not as anything that displaces people. If you think about this, at the dawn of the pandemic, we had record low unemployment, but you would have said that we were at the top of what humanity had as technologies. And so why was unemployment so low, then? Well, that’s because all companies, you know, need to generate higher levels of productivity. And so if you think ahead for the next 20 years, we’re not going to be able to increase the size of the human population by 10x. In order to have 10x more productivity, we’re going to need machines to help us.
WD: So, a strong AI like Pryon can actually be a complement rather than a threat to the workforce.
IJ: Absolutely. When we show our platform to folks, you can see them breathe a sigh of relief…the fact that they can immediately save up to two hours per day per employee. That’s not time that’s wasted, that’s time that they can actually work on more complex problems. So it’s a win-win for everybody.
WD: Terms like AI, machine learning, natural language processing get tossed around so liberally in the startup world, what do you think makes an AI solution legitimate?
IJ: I know a lot of startups just put.ai in their URLs and suddenly they think they know what they’re doing. I don’t know how to describe their legitimacy, I can tell you what describes ours: We build our own engine stack, the whole thing, soup to nuts. Because when you think of how Apple tends to solve problems in our personal lives, the fact that they can build the silicon all the way up to the operating system all the way up to the applications that they host….this gives you the highest level of performance with the lowest possible power use as well. And obviously, we all need to be more thoughtful about being as green as possible. And so same thing here, right, how do I get the highest level of performance vis-a-vis accuracy, while at the same time, being able to meet the enterprise’s conditions to thoughtfully operate in their environment.
For me, what makes a real AI company is your ability to make your own engines, not repurposing open source, not being a glorified integrator of things that exist in public cloud. You can strategically leverage some that, but there has to be something unique that you can only buy from me.
WD: I think that makes sense. And it’s a great segue to enterprise technology. I heard you on our recent two minute podcast and you mentioned that Brian is taking a multi cloud approach. Why do you think multi cloud is a good fit for startups?
IJ: Part of it is out of necessity; there just isn’t enough resource to go around. All businesses are essentially reinventing themselves with AI. So the way that we leverage OCI is for training models that then get deployed into production, whether it’s entity extraction, optical character recognition, and the like, scientists have to find spare GPU capacity and then leverage it to develop these models. It seems heavy on the compute side, but then you think about how much time we save for these workers, once these models are in production, and it actually has a net reduction in carbon.
WD: Well, that’s a win win. I was going to ask where OCI (Oracle Cloud Infrastructure) fits in. And if there’s certain workloads that shine with Oracle Cloud, it sounds like model training, high performance compute is where it’s shining for Pryon.
IJ: That’s exactly right. And I think Oracle’s a leader in moving to cloud applications, whereas if you look at AWS, for instance, they may have a lot of the infrastructure components but they don’t have a lot of applications. And obviously Oracle Applications are prevalent throughout corporate America and then certainly internationally as well. So at the right time when we have to draw in more application centric data for the for the AI to essentially leverage as part of its knowledge management workflows, that is going to be pretty important.
WD: And that takes a lot of cooperation. And speaking of which, Brian is a member of the Oracle startup program. How has that experience been? Have there been any highlights?
IJ: The highlight is the team’s willingness to explore how we can work together even beyond the normal relationship that you would expect between a startup and a host company such as Oracle, and then for Oracle to leverage these types of technologies for its own workloads as well becomes a self-fulfilling prophecy. Consider a big tech company as a patron of essentially outsourceed R&D, which is exactly what the startups are in these programs, and then when one has a hit, and is relevant to the rest of their clients, Oracle can prove it out for themselves and then essentially run campaigns to say, hey, this helped our business in this particular area. Then we can rinse repeat this to every one of our other clients and it becomes a self-fulfilling prophecy.
WD: Yes, we call that a virtuous cycle of innovation. So, Pryon raised a major funding round this year, first of all, congratulations, what other exciting things happening for your company?
IJ: We had a strategic access round, in which we were able to get more of our customers to participate in the growth of the company as well. We actually had an Oracle veteran join us as our chief revenue officer who’s also on our board now, Chris Mahl, it’s been a delight to work with him. We’ve had Kirsten Wolberg join our board, she was former CTO of DocuSign, and former CIO of Salesforce. So, we’re building up our go-to-market presence because our R&D and engineering have built a platform that’s leading to market-leading levels of accuracy. Now we have to build a sales and marketing team to get that absorbed by the market. So, we have some exciting announcements that will be starting in the first quarter timeframe for some new customers and the next version of the platform that we’ve been working on in the past year.
WD: Igor, I’m really excited that you chose to stop by.
IJ: Thanks for having me.