Benzinga’s Fintech Focus Podcast features conversations with the biggest names in fintech. Subscribe to the Fintech Focus newsletter to get a roundup of industry news delivered to your inbox weekly, and check out upcoming programming at Benzinga events.
In this episode of the Fintech Focus podcast, we talk about the science of, well, talking.
Conversational AI has already changed how we interact with computers—thanks in large part to Siri, Alexa and Google Assistant—but we’re only just beginning to see the transformative power that this technology can have on our daily lives.
Leading the charge in this arena is Clinc, a company that was born out of a research project at the University of Michigan.
On Episode 27 of the Fintech Focus Podcast we chat with Dr. Johann Hauswald, co-founder and chief product officer of Clinc, about the company’s beginnings, why they’ve chosen to focus on financial services right now, and what makes conversational AI such a difficult technology to crack.
Listen to the podcast below to hear how Hauswald is helping to help out the financial services world through the help of artificial intelligence.
Listen to the podcast below to hear how Hauswald is helping to help out the financial services world through the help of artificial intelligence.
Speaker 1: 00:02 Welcome to Benzinga’s Fintech Focus Podcast, your weekly dive into the world of Fintech featuring conversations with leaders on the front lines of the financial revolution.
Spencer Israel: 00:18 Hey everyone. Welcome to episode 27 of the Fintech Focus Podcast. I’m Spencer Israel. When most of us think conversational AI or voice assistance, we probably think of one of two companies, Apple, which launched Siri in 2011 or Amazon [inaudible 00:00:35] with Alexa in 2014. But there are other players in this space, and today we are joined by one of them and they are Clinc, a conversational AI company that was founded by a group of researchers at the University of Michigan in 2015.
What happened was the published a paper that essentially open sourced conversational AI, and it got such a good reaction that they decided to start a company and that led them to Finovate Fall in 2016 where at their presentation won best in show.
Now Clinc’s a little bit different for us because they’re not really a financial services company, but that’s the vertical that they focused on for now, and they can already count some of the biggest firms in the world as clients including Barclays, S&P Global and USAA. I was able to catch up with one of their co-founders, Dr. Johann Hauswald, Clinc’s Chief Product Officer. So, without further ado, here’s my chat with Dr. Hauswald from Clinc. Johann, thank you so much for joining the Fintech Focus Podcast today.
Johann Hauswald: 01:37 Happy to be here. Thanks for having me.
Spencer Israel: 01:39 I like to start off these interviews just by getting a quick background for the record about how you got where you are, how you got interested, generally speaking in financial services, but in your case it’s more about the AI as a whole. So why don’t you explain your background and how you ended up co-founding Clinc?
Johann Hauswald: 01:59 Definitely. So, I started as a co-founder at [inaudible 00:02:05] company in 2015, and before that I was a graduate student at the University of Michigan studying computer science and doing my PhD.
I met my co-founders… Actually two of them are my PhD advisors, and one was a fellow grad student at the university, and I met him at the research lab where our focus, and clarity lab, at the university was really in bringing the best science and technology out and doing research as to the impact that new technologies are going to have on future infrastructures of the world. And so, in our case we were looking at machine learning applications and how data centers would need to change in the future for these new technologies that are voice image, natural language processing based. And so, through that research we published a research paper called Serious where we created the world’s first open source, intelligent personal assistant.
And really the goal of this research of what’s to understand how do you build an end to end conversational AI system? Because at the time, really the only company that had access to this knowledge where the Apples and the Googles of the world, this was back in 2013 and 2014 where this technology was not freely available. And so, our goal and our mission with this research was to democratize this knowledge. And in publishing that research, it had massive impact in the field. We published a top tier paper in early 2015, and there was just immediate buzz right after that. We had tons of interest from the market to actually commercialize this technology across many different domains. So, I remember there was a law enforcement officer in Australia that reached out to me and said, hey, this technology would be great in order to search to search these massive databases that we have.
Medical, automotive, financial spaces companies all reached out to us asking, it would be great if you guys could take this technology further, because we would like to use this for applications internally.
I remember a distinct quote from an interested enterprise who said, “We want it to be our technology in the car and not have it be licensed by Siri or Apple for example. They didn’t want Siri for their car. They wanted it to be the Ford and that experience for them in the car. And so, on the heels of that research, we founded a company in July of 2015 and so I was … Part of the goal and the mission was really to take some of the research that we did and commercialize it. And the company’s mission is really to use conversational AI to enable the access to knowledge. Through this experience, you can really provide people more information than what they would traditionally have what’s available to them today. If you look at our focus in the financial space, what we’re doing is taking information that would otherwise take a banker or somebody on the other side, potentially minutes or hours or days to even collect. And we’re making this readily available through conversation and through natural language.
And so, our interest to start in the financial space really stemmed from looking at the different verticals and identifying that in finance or even need for that democratization. There is a human necessity to bring people closer to their finances to actually solve a real problem and to improve their quality of life. That’s part of the mission, which is improve access to knowledge. And by doing so, improve the quality of life to then be more financially responsible and aware of one’s financial story. So that’s the overall story how specifically we found the company and chose to focus on finance first.
Spencer Israel: 06:01 Serious you mentioned was open source. Did you hear from Google and Amazon at the time, or “God damn it, that was our thing?”
Johann Hauswald: 06:12 Interesting question. And yes, we call it Serious to be a little play on the wording, but it’s also a star in the constellation. And we did get a letter from the general counsel from Apple a year later, or it must’ve been over winter of 2015, saying, “We wouldn’t want there to be any confusions as to where what you guys have built versus what our commercial product [inaudible 00:06:41] for. So, they didn’t really threaten to sue or anything, but they used some strong language.
And at the time, we had just created, this was an academic paper we had just kind of gotten around the idea of starting a company, and it wouldn’t look good for Apple to sue the university that was open sourcing this technology. But in creating a company we didn’t want to have some negative press to start with. And so, we renamed the project [Lucida 00:07:16] in the open source community, but the academic paper is still called Serious to this day.
Spencer Israel: 07:22 I would imagine if you get a letter from Apple’s lawyers, you’re either doing something incredibly right or incredibly wrong. There’s really no in between there.
Johann Hauswald: 07:30 Exactly. And the university actually wanted to … They rushed to ask us, “Change the name, change the name,” and we stood firm and said, “No, we’re not changing the name.” And for a while that’s how it stuck. And then we did update, as I mentioned it, the open source one, but apple didn’t really want to mess with the academic community because that would set a bad precedent for them hiring researchers, especially in the field of machine learning. So, I think they made a wide decision there to let us retain the name on the academic paper.
Spencer Israel: 08:04 So Clinc has, I think four co-founders. All of you were working together at Michigan?
Johann Hauswald: 08:11 Yes, that’s correct. So, two founders, two of them are professors at the University of Michigan. So, Jason Mars and Lingjia Tang are Assistant Professors, and Mike Laurenzano, who is the CTO was with a graduate student as well that is a coauthor on the Serious paper.
Spencer Israel: 08:31 And so what were those early conversations like? You wanted to explore the commercial applications of this technology, but how were we going to go about doing that?
Johann Hauswald: 08:43 Yeah, that’s a great question. In building this technology, on one side we were academics that we’re looking to understand… The research was about understanding the complexities of managing this type of new emerging application. The challenge when you have such a complicated software stack is in the processing and making sure that you can deliver the experience in a way that’ll be useful for users if you use the best machine learning, but it takes 10 minutes to respond. No one’s going to use it.
And so those early conversations were really around, how can we make this software solve an actual problem? And so what we initially thought was important in terms of how quickly we could answer the questions, we also quickly realized that in order to solve a real problem and to, as I mentioned, for example, provide a user an accurate picture of their financial stories, we had to innovate and create some very advanced machine learning tools and innovate on the algorithms that were stated in the arts in academia at the time of creation. And we’re still innovating.
And so really what we focused on, what is the end to end experience that a customer would have if they were to just talk to this technology. We quickly identified early on that a lot of technology in this space didn’t satisfy the needs and the expectations that people had of conversational technology. There’s a lot of hype around Chatbots and being able to talk to AI, but nobody could point to an experience that says, hey, this experience I trusted, and I can really just have a conversation as if I’m just talking to somebody in the room. We like to use the phrase, “Human in the room level understanding” where you’re really just looking to converse with an AI as close as possible to this just being your banker or in a different case, it could be your doctor for doing a medical application that is conversationally enabled.
Thinking of the final goal of providing an awesome experience that solved a problem end to end where the information was presented to the user in a way that informed them through, as I mentioned, our mission, really forced us and challenged us to innovate in this space across the stack. Not in machine learning, but then also in how, in the what we call human computer interaction, the HCI aspect of this where we had to, in order to present what a user was asking, we had to serve them up with some content. The first time we unveiled this in 2016 it was groundbreaking at Finovate Fall where the experience that we showed was really end to end where somebody could just talk and then receive insights into their finances. And that really drove what the company is today, which is how to use conversational AI as a tool to solve real world problems and challenges that people are having.
Spencer Israel: 11:50 So explain then how the technology works without getting too into the weeds and too specific. Obviously, you don’t want to give away the secret sauce, but like you mentioned, there are conversational AI platforms out there. Siri’s not new. Alexa’s not that new. It’s a couple years older by now. How is Clinc different?
Johann Hauswald: 12:12 Sure. So, the fundamental approach that we take is to really push the technology to, as I mentioned, to solve a problem and to be robust to messy language. The first challenge that we took on early to do was to tackle the problem of voice, and voice, it’s messy. People say many things. They may stop, stutter, say, oh actually I didn’t mean that. And so, voice was a good playground for us to really push the bounds of the technology, and it made us innovate in how we’re [inaudible 00:12:50] these machine learning algorithms. Part of our process to build this conversational AI, we early on identified collecting crowdsourcing data. So, data from the crowd that represents how you would talk to an AI as a fundamental piece in the pipeline of building the technology.
So, what we do and what we recommend to all of our customers to do in thinking about building conversational AI is to collect data that most closely represents the use case that they’re looking to solve. And to not worry about how messy the data is because the messiness of that data represents how people would actually talk. And so, our algorithms are architected in a way to support that messy language that’s unconstrained where we don’t need to put parameters over how a user or a customer of this technology would talk. But if you have algorithms that are robust and advanced enough where the data can be messy and have variability in it, then you have an awesome experience that just allows people to just talk freely without thinking. And so, we’re using
Johann Hauswald: 14:00 … using the latest science and technology in machine learning, deep learning algorithms that are data driven. So that once you’ve built the experience maintaining it is really about adding more training data to expand the knowledge of the AI as a day-to-day upkeep of the system.
Spencer Israel: 14:20 Okay. So, people are gonna talk how they’re gonna talk. But what are then the biggest challenges in making this technology? It is still in its infancy.
Johann Hauswald: 14:31 Mm-hmm (affirmative). So, I’d say the biggest challenge is really making sure that the expectations, and this is more of … I wouldn’t call it a challenge of the technology, but more of a challenge of designing the experience, is really setting the expectation of what the system can and cannot understand. Because once you’ve set that expectation early on with a customer and with somebody interacting with conversational AI, then they know to interact with it. And the way that we’ve addressed and we’re solving this problem is by doing a lot of upfront work of scoping and designing the experience. And so, we’ve coined some new terminologies in this space where we like to use … typically folks use the word intents to define what a system does and doesn’t understand, and we use competencies. And what that has created is a way for us to reason about what is the AI knowledgeable about and what isn’t it knowledgeable about.
So, a competency can be something about … your AI knows about balances or transactions. Those are very broad topics, but if I hand you my phone and say, “Hey this experience,” … I have a conversational AI that is liked to your bank account and you can ask it anything about your balance or your transactions, then the promise that I’m giving you is that no matter which way you ask it, whether you ask transactions at a certain location, on a certain merchant or on a certain date, you should be able to ask that question and it should be able to support that answer.
And so really setting the expectation of users of conversational AI early on is … has been the … one of the challenges that I say we’ve addressed and that we push all of our customers that we work with to really define up front. It’s kind of like when you go to Alexa, I get … In talking to people about conversational AI there’s always frustrations about, ” I don’t know what I can ask?” But if you’ve framed the experience of, “Hey, I’m your conversational AI on your bank account. You can ask me about your balance, or about transactions, or just get advice on spending,” then you kind of know the realm of what’s possible. And the promise though, in the design of that experience, is that it can handle everything there … that a customer could potentially ask about in balance or transactions or spending advice.
Spencer Israel: 17:00 We talk a lot about a fintech on this podcast obviously, but I wanted to take a quick second to tell you about our fintech product, Benzinga Pro. It’s Benzinga’s real-time news platform, great for anybody who trades using news in the markets. With Benzinga Pro you get access to breaking news headlines, press releases SEC filings, Audio Squawk, chat rooms and a whole lot more. Now normally, I would say go to pro.benzinga.com to get a free two-week trial but, because you listen to this podcast, I got the hook-up.
If you go to that site, pro.benzinga.com, enter the promo code focus, as in Fintech Focus, you’ll get 25% off a monthly or annual subscription to our Essential package. It is our most popular package on Benzinga Pro. So again that’s pro.benzinga.com, promo code focus to get 25% off.
And now, back to the show.
So, you were studying this as far back as 2013. How long did it take you to have a product that you could sell?
Johann Hauswald: 18:09 So we founded the company … We started studying this in 2013, but we really started building a productizable offering in January of 2016. So, we founded the company July 2015. For a couple of months, we were studying what we wanted to focus on, especially what space we wanted to attack. Because also in what I mentioned in setting the expectation of the users, there is a conversational AI. The way it can really work well is if it’s scoped correctly. And so, we wanted to attack a certain topic area and vertical. So, finance was the one we started on first.
And so, it took us eight months initially to build that offering that we presented at a financial, at a fintech conference [inaudible 00:19:02] 2016. So, we built from January 2016 to September 2016 where we had a small team of researchers from the university. So, in addition to the fore founders of the team, other PhD students early on joined to be part of building that initial conversational AI stack, and so really a lot of the fundamentals and research and technology that we built were done in those eight months. And that has been really the foundation for our technology. After … Now what we’re doing is basically platformizing that technology. So, a lot of the … sort of the back-end machine learning algorithms that were built initially are now going into a platform that we’re providing our customers. But really the first product that was built initially in those eight months was the first iteration of it.
Spencer Israel: 19:54 At any point in those eight months, did you ask yourself, “Are we gonna be able to convince financial firms to use this?” Was that part of your discovery process?
Johann Hauswald: 20:08 Yeah, that’s a great question. So, we had investors at the time that were kind of pushing us in that direction, to do some of that market validation along the way, but that’s the work that we did before we started building. So, our approach was, “Let’s get a sense for what the market wants,” and we had gotten that market validation early. The moment we published the paper in March 2015, and I gave the academic talk in Istanbul, in Turkey, our … the technology, the open source project went up on Hacker News and it blew up. There was GitHub. There was … it was one of the top trending, top ten trending projects at the time. And so, we had a lot of validation initially that this technology was something that people were interested in.
The use case … we chose financial services and we started to have conversations. We had one enterprise at the time that was really interested and so we knew that they were some folks that we could continue having conversations with. But between January and September when we built the experience, we really trusted ourselves that if we were proud of this product, and that we thought it would be something that would change our lives, then it could change other peoples’ lives. And so concretely what that would … the way we approached that was really to study, like what are the initial topics that we would want to cover because it’s a big endeavor to say, “Hey, we’re gonna tackle the financial space,” but we wanted to have something out in market relatively quickly and so in eight months we said, “Okay, let’s have a set of topics that we really can dive deep into and that people can just trust the experience.”
And so, the initial experience that we called [Fenie 00:22:00] for the financial genie, had about 10-15 of these competencies that spanned an experience that you could just integrate into your mobile app and you could just talk to your bank accounts. As I mentioned eight support things as like balances, income, when did I last get paid, how much money can I spend, show me some information about my transactions.
Spencer Israel: 22:22 Okay. So, you knew then that there was an appetite for it.
Johann Hauswald: 22:25 Definitely.
Spencer Israel: 22:28 Okay. Good. And then before we dive too much further into this vertical, what other verticals are you guys in right now?
Johann Hauswald: 22:35 So we’re currently … The financial space is the one that we started in so that’s our most actively … that’s where we have the most active partnerships. Quick service restaurants, QSRs, there’s a lot of interest there. Medical, automotive and I think … and gaming. Yeah, we just … We were at a Gaming Developer Conference, GDC in San Francisco a couple weeks ago where we unveiled a prototype where you can have an in-game conversational experience.
So it was a really exciting demo where, and we’ve gotten a lot of traction since … where you are a character in a world and you go up to a shopkeeper and you need to trade or buy a new weapon and the shopkeeper has so many weapons and instead of having to point, click and scroll, you can just say, “Hey, I need a sword that is good against an ice character and that has at least a damage ratio of ten or something.” And so those conversational experiences could really augment what currently exists, or what doesn’t exist out there in the market.
We had … Our team was saying that there was nobody in the gaming space right now that’s thinking about this, so it’s pretty exciting.
Spencer Israel: 23:45 And looking at your roster of clients in the financial space, it sort of reads like a who’s who. You mentioned you’re in Turkey, you’re in Isbank there. You’re in USAA, Barclays, S&P Global, U.S. Bank. What has the process been like of pitching these banks, or these financial firms, and how receptive have they been? And walk us through sort of how you’re going into the relationship with financial firms and what the reaction has been once you talk to them to this technology.
Johann Hauswald: 24:19 Sure. So, a lot of … What’s been interesting is there are some key players, especially our early adopters, that have been hoping that there is a technology that can come and disrupt in the space. So, for example, USAA, they’ve been looking to conversationally enable their enterprise for a while. And then they were one of our earliest customers and so it’s been very straightforward I’d say with the customers that have been thinking about this because they know what the technology out … is out there in the market, and they’ve done the validation to say … and when they see our experience they say, “I haven’t seen anything like this.” Because basically what we can do is we say, “Hey we’re a conversational AI technology. We’re not gonna bore you with a set of slides, we’re gonna show an experience.” Because at the end of the day, and this is back to what I was saying early on, as to how fundamentally we think about this, it’s about the experience that a customer will have. And so, when we show the experience, we get a lot of interest and buzz where they say, “I haven’t seen anything like this where people can just talk to this AI.” You know its typically very scripted, very single turn demos, but here we train our sales and product team to just, almost do unscripted demos of the experience to show just how natural the experience is.
I remember when we went to the U.K. and did a demo day with Barclays, they were floored. They had done three others … They had done this massive RFP of 30 vendors. They had an incumbent at the time. And they were … You could tell that it was the end of the day, they were like, “Oh, who’s this startup from the U.S. showing up here?” And we did a demo with all of our different channels, so we had an Alexa in there, we had our web demo where you could just talk to it, and they were floored. They were tired from the day and they just saw it and they were like, “Wow, we have not seen such an exciting and just fluid and easy demo to talk to.” We also had some of their people come up and play with it.
And so that has allowed us to use this approach of show, don’t tell. And then what the banks and what the enterprise typically goes through a phase of proof of concept. Like, “Hey, let’s integrate this internally with our services.” For the case of Fenie and that financial experience, we architected it in such a way that we can just conversationally enable their bank accounts. All we need is APIs for transactions and balances, and that’s a very rapid process. In a couple of weeks, we get integrated and we can show … You can actually be talking to your bank account, which is an awesome experience. And then they go through some larger pilots where we had a customer that just … after a couple of days, and this has happened multiple times, USAA included, where after a week they say, “This is awesome.” They integrate it into their APIs and all of the internal folks that are playing around with it say, “Wow, I haven’t been able to get this information from my bank account ever. How much did I spend last week?” Typically, right now you have to go to your bank account, you sum it up, but here they’re just like, “Oh this thing can just tell me how much I spent and then I can ask it just about food or certain categories.” And so that kind of sequence of events then leads to them wanting to get deeper in the relationship and build some additional use cases. And for that, since founding the company, we now have a platform that enables enterprises to build and adapt and customize the conversational AIs so they’re not just starting with our use case that we initially built, but they can build
Johann Hauswald: 28:00 -any use cases that they want, so for Barclays for example they chose to have something on Facebook Messenger all around help and support topics. And so we spent a couple of months building with them conversational experiences where their customers can go to Facebook Messenger and ask about fraud or how can they change the details on their bank accounts, like their address, their pin, how they can make international wire transfers, things like that. And so, the use cases end up being very enterprise specific, but for that reason we have a platform that enables either the enterprise or partners that we work with to build these custom use cases for all these enterprises.
Spencer Israel: 28:39 So who is building that? Is it Barclays or you?
Johann Hauswald: 28:42 It depends on the model, so-
Spencer Israel: 28:43 Like for each customization there?
Johann Hauswald: 28:47 Yeah. Right now, a lot of the experiences are being built by the customer. We’re moving into an advisory role, but there is a period of time where we’re co-creating because as I mentioned our technology works very differently than what’s out there in the market. And so, we’re building together with them and [Barco 00:29:12] was actually here for two weeks a week ago with ten people from offices in India and in the U.K. to build conversational AI, to learn how to build conversational AI.
And so really what we’re seeing in the market is the large enterprises really want to equip themselves to understand and build so that then once they have the platform internally they can train other lines of businesses to build conversational AI and then to target those high value use cases that will lead to cost reductions internally. One of those cost reductions is building so that no one’s calling on the call center and then you don’t need to grow the call center. And so really, they’re looking to conversationally enable their enterprises.
Spencer Israel: 30:00 Do you remember the first firm you ever pitched Clinc to?
Johann Hauswald: 30:07 The first firm, so USAA was our first customer that signed on. The first firm, I mean, it must have been at Finovate… So, I know Wells Fargo was one of the first, when I mentioned enterprises being interested in Clinc right after the research, that was Wells Fargo, they’ve been dabbling in the space for a very long time. And so, they came, and they had an innovation that had been investigated in the space for a while. We had early discussions with them.
But the first firm we formally pitched… I mean, basically when we did that unveiling at Finovate 2016 we pitched all across the board. We specifically targeted that event as it’s one of two of the largest fintech events where we could pitch top executives in the banks, so I would say we went across the board and we wanted to go after really the largest banks because that would enable us to get our technology out there with the biggest customers. Barclays, they have 45 million customers total and so one of the missions and especially missions of the founders to make sure that we’re having high impact and high impact is through getting into the top ten banks of the U.S. and of the world.
Spencer Israel: 31:35 What was the initial reaction like in the few weeks and months after Finovate?
Johann Hauswald: 31:43 We had a lot, a lot of interest. It was surprising. I mean, we knew that people would be very interested, but we didn’t know to what scale. We actually won best in show at that event and we subsequently won a decent number of awards at these trade shows because the experience that we show, it’s just not something that people can find in the market, and so there was tons of interest. I remember being on the phone, back at that point we were a relatively small team, and so then our engineers are… we’re basically doing sales calls, customer success, troubleshooting, all at the same time with such a small team. But we were really shocked and excited as to the reception. Maybe not shocked, because we knew that this technology was very different than what anybody had. But you can only prepare so much for how excited people are going to be about your technology.
Spencer Israel: 32:47 I believe you last raised money in 2017, is that something that is on your radar going forward or are you guys satisfied where you’re at right now in terms of capital risk?
Johann Hauswald: 32:58 Yes, so we did our [inaudible 00:33:02] around early 2017. We’re actually in the midst of raising right now, so two years later we’re… we actually signed a term sheet today for a [inaudible 00:33:14], so that’s moving along, so we’re in the middle of those discussions. We’re actively excited to grow and scale the company, to keep conversational enabling, not just the financial space but across many, many verticals.
Spencer Israel: 33:34 What would you say is the biggest issue right now, the biggest challenge facing Clinc?
Johann Hauswald: 33:42 Good question. We have this very powerful technology and really one of my cores focuses and I think one of the companies is enabling these enterprises to build for themselves. We know how to build conversational AI and we know the level, the bar, at which it should be built, but we need to innovate in a way that we train and educate enterprises to build a lot of… We did a customer summit two weeks ago where we brought together our top customers and one of the prevailing themes there was, and this was actually voiced by Barclays, they presented to others, was that the training and knowing how with this technology is extremely important in order to conversationally enable your company.
And so, we’re heads down focused on creating an offering that makes sure that customers understand not just the how of using our platform, but really the why and the dark magic that goes behind building robust and advanced conversational AI.
And I think another challenge, but it’s a welcomed one, is just attacking the next vertical. We know that this technology is vertical agnostic. We proved that hypothesis a while ago when we went from building in fintech to… well, one building in a different language, even though it’s not a different vertical, you need to make sure that your technology is pretty agnostic to the data that you feed to it. But then when we also started working with Ford, with whom we have an active partnership with, and that was announced mid or late last year, we’re going from building in finance to building in automotive, that means that our platform is vertical agnostic. And so then finding all the places where we can conversationally enable in a fun and exciting new development and challenge for the company.
Spencer Israel: 35:49 Is it hard to explain how this stuff works to non-computer scientists?
Johann Hauswald: 35:58 Well, the thing is most of the time you don’t necessarily need to explain how it works because you show how it works. So if you mean by how it works, the technology, not necessarily because the way that you use machine learning, it’s to solve and so if you know the fundamental application of the technology, you don’t need to dive into the theoretical background as to how a machine learning algorithm works. There’s a practical application of the algorithm.
So, if I read a research paper in order to understand how we can solve a problem, what I’m trying to understand from it is what kind of problem can this solve for me and so that’s the way I explain it. You have different problems that you need to solve along the way of providing and of doing natural language understanding, and so then decomposing it into sub-problems that then each has a tool to accomplish that problem. It’s typically the process that I follow, which I think resonates with folks.
Spencer Israel: 37:03 I like to wrap all these up with the same question and that is what keeps you up at night?
Johann Hauswald: 37:09 I want to make sure like I said that what our customers are building and what we’re building together for conversational AI really maps back to that overall mission, which is to enable and facilitate access to knowledge and that’s a very exciting challenge that we take on, but it’s not without its stop-and-starts on the way. Because some people might want to use conversational AI for different reasons, right? Our fundamental reason is to use the best science and technology to enable access to knowledge and if there’s a different use case and it’s not one that we’re in line with, then that keeps me up and I want to make sure to guide customers.
Because this is really powerful technology that can be used for so much good and if there is some cases where it’s used in, I won’t say improper ways but in ways that don’t do the technology justice, then guiding customers and saying, “Well, hey, you could think about it this way and you could solve such a larger problem,” or it could have even more impact is one of the challenges that requires a whole company to band together. It’s not just the engineering team that’s says it, but it has to be our whole entire team that really pushes this message out in the market that conversational AI is a very powerful tool that you can use to definitely change peoples’ lives.
Spencer Israel: 38:36 Dr. Johann Hauswald, thank you so much for joining the Fintech Focus podcast.
Johann Hauswald: 38:40 Thank you so much for having me.
Spencer Israel: 38:44 I want to thank Dr. Hauswald for being very generous with his time and thank you of course for listening. If you enjoyed this interview, please leave us a review on whatever platform you’re getting this podcast right now. You can also Tweet at me @sgisrael or email me spencerisrael@ benzinga.com
Also, don’t forget to subscribe to our newsletter. The Fintech Focus newsletter goes out every week. You can get that by going to benzinga.com, clicking on the right-hand side of the page.
Also, before we go, I want to tell you about the Benzinga Training Summit, June 20th in New York City. If you’re an active trader or investor, then this is the event for you. Featured speakers include everybody from ZNBC, John and Jerrian and Tim Seymour to Chris Irons aka Quoth the Raven on Twitter to proprietary traders who will walk through their strategies and everybody in between. And if that’s not enough a portion of the proceeds from this event will go to Traders for a Cause, they’re a non-profit that helps traders give back.
Now Fintech Focus listeners can get a discount by entering the promo code focus15, that’s F-O-C-U-S, 15. But we are limiting tickets to account for space, so to get your ticket and learn more go to benzingatradingsummit.com
Thanks a lot, and we’ll see you next week.