John Sun, CEO and Co-Founder of Spring Labs, on building products (and an event!) for AI in financial services

Enjoying the podcast? Don’t miss out on future episodes! Please hit that subscribe button on Apple, Spotify, or your favorite podcast platform to stay updated with our latest content. Thank you for your support!

John Sun, Co-Founder and CEO of Spring Labs

“AI is probably the most talked about and least understood topic within financial services today,” said John Sun, the CEO and Co-Founder of Spring Labs (previously, I had another co-founder of Spring Labs, Adam Jiwan, on the show back in 2020). The company is at the cutting edge of AI innovation in financial services with several interesting use cases in production today.

The Spring Labs team decided that what the industry needed was a new event focused on AI applications beyond just the use of chatbots. So, they created the AI-Native Banking and Fintech Conference, happening in Utah on October 7. Tickets will be available here right up until the show kicks off.

In this podcast you will learn:

  • How Spring Labs has evolved since its founding in 2018.
  • Where they are focused today.
  • What Gen AI applications they have in market today.
  • The key markets they are focused on.
  • The first use cases they deployed for their Gen AI product.
  • How they are inserting AI tools into existing workflows.
  • The outputs that come from these tools.
  • Why they decided to start an AI-focused fintech conference.
  • Why they chose Salt Lake City as the location for this first event.
  • Who is attending the event.
  • The AI topics they will be covering.
  • What John is personally looking forward to on the agenda.
  • How people can get tickets to the event.
  • John’s vision for Spring Labs.

Read a transcription of our conversation below.

FINTECH ONE-ON-ONE PODCAST NO. 502 – JOHN SUN

Peter Renton: Welcome to the Fintech One-on-One podcast. This is Peter Renton, Co-Founder of Fintech Nexus and now the CEO of the fintech consulting company Renton & Co. I’ve been doing this show since 2013, which makes this the longest running one-on-one interview show in all of fintech. Thank you so much for joining me on this journey. Now let’s get on with the show.

Today on the show, I am delighted to welcome John Sun. He is the CEO and Co-Founder of Spring Labs. Now, Spring Labs is a super interesting company, been on the cutting edge of technology in fintech for many years. But I wanted to get John on the show now because I think what they’re doing right now is one of the most interesting applications or programs for generative AI that I’ve come across. We talk about what they’re doing when it comes to helping frontline agents with complaint monitoring, and there are many different features of this product, which we do go into in some depth. We also talked about a new event Spring Labs is hosting that is coming up quickly on October 7th, the AI Native Banking and Fintech Conference. John talks quite a bit about that, why they decided to do it, who it’s for, and what they will cover on the day and why people should attend. It really was a fascinating conversation. Hope you enjoy the show. Welcome to the podcast, John.

John Sun: Thank you, Peter. Appreciate you having me. Excited to have a chat about this.

Peter Renton: Sure. My pleasure. So let’s kick it off by giving listeners a little bit of background about yourself. I know you’ve been in the fintech space for a long time, but maybe you can talk about some of the highlights of your career to date before Spring Labs.

John Sun: Yeah, absolutely. And again, thank you for having me, Peter. You have definitely been in the fintech space for quite a while. I think I was at one of the first LendIts you hosted way back in the day. But really, my background is I started my career in fintech, running analytics and online retail finance for Enova. And this was kind of the very early days of online lending, the very early days of what we call fintech today. From there, I spent a couple of years in San Francisco building fintech products with Y Combinator from 2011 to 2012. And ultimately kind of wound down that business and started a lender in Chicago with my co -founders, Al and Paul called Avant. And Avant was one of the early pioneers, I would say, in the use of modern underwriting techniques for online lenders. We’ve done a lot of work very early on with machine learning techniques and using those in real time customer underwriting. And I think that’s ultimately what gave us a bit of a competitive advantage over other lenders of the time and really was the core of what made Avant different in its era. And in 2018, I left Avant to start Spring Labs. And the original mission of Spring Labs was to really productize the next generation of technologies that’s likely to be transformative within financial services and make those technologies easy for folks to adopt and really use to benefit their customers.

Peter Renton: Right, yeah. And, you know, I had your fellow co-founder, Adam Jiwan, on back in 2020. I’ll link to that in the show notes where you were talking about building sort of a credit bureau of the future, or at least building sort of the data infrastructure for it. Tell us a little bit about the history of Spring Labs and how it has evolved over time.

John Sun: Yeah. Funny you bring that up. It’s been quite a while since we were, you know, really a hundred percent focused on that particular piece of it. And I think really the history there is when I started Spring Labs, the goal again was to figure out what are the next technologies that are going to be transformative, you know, for financial services. We weren’t super opinionated on which ones those were. At the time, it looked like blockchain technology was going to gain widespread institutional adoption, really powering the next generation of fintech startups. And that’s where we had gone into it with the attitude of, well, where can we apply this technology to really interesting problems? Data exchange was one of these interesting problems where it was difficult for existing financial institutions to exchange data and to do it in such a way that you can preserve the anonymity of the sharers and preserve the privacy of the customers that you’re sharing data about. And that was really the first kind of iteration, I would call Spring Labs 1.0, which was really about using the most cutting-edge technologies to enable that data exchange. Ultimately, we found that institutional adoption of blockchain technology was a little slower than we would have liked. So we started really looking for what are the other technologies on the horizon that are likely to be transformative. And that’s where in kind of 2020 and 2021, we really started experimenting with the earliest versions of generative AI and looking for opportunities to apply generative AI within financial services. And today, that’s the majority of our business is building easily adoptable products and easily implementable generative AI products for financial institutions.

Peter Renton: Okay. Well, personally, I think it’s a shame that first product didn’t work out. It was groundbreaking, revolutionary even at the time, but you might’ve just been a little bit ahead of your time there. But can you dive into the product suite today? You talked about Generative AI, but what applications do you have in the market right now?

John Sun: Yeah, yeah, absolutely. And it’s funny you bring up that, you know, maybe we’re a little ahead of our time for the data exchange product. I still really believe in that product as well. I think there’s a lot of value in that. You know, funny enough, we are still kind of running and maintaining a couple of different data exchanges using that technology. And I think the users are getting a lot of value out of that. In terms of the generative AI products, there’s a couple of areas where, you know, I really wanted to focus the team. I think number one was compliance. There are a lot of opportunities for the unique value that generative AI technology brings to the compliance sphere. And I think the other is what we call conversational workflows or conversational agents where essentially we’re able to convert the tools that you’re already giving your customers or your agents in the form of web or mobile applications and adapt those to conversational interfaces. Essentially, allowing your users, agents, or employees to interface with those tools by talking to a chatbot rather than navigating a website or a web application. And I think there’s obviously a lot of unique value that generative AI also brings to that capability.

Peter Renton: Are you primarily focused on financial services for that? Because when you talk about it like that, that could be just about anything. But how are you applying that to financial services?

John Sun: We are almost 100% focused on financial services. And I think the reason for that is I believe vertical specialization is one of the only defensible competitive advantages in this AI field. I think there are going to be a ton of AI companies doing a ton of very similar things. But really, what unique advantage you can bring is to understand your vertical well and to bring a lot of the unique efficiencies and learnings that you would only have insights into within a particular kind of vertical. So, as a result, we only work in financial services. I think the other side of it as most of your listeners probably know is that building products for financial services is very different from building products for say e-commerce. A lot more regulations, a lot more requirements for accuracy, a lot more basically that needs to be done. So as a result, we really wanted the tool, our kind of toolkit to be the right solution for financial services clients.

Peter Renton: How are your clients using the product? Are they augmenting their existing operation? Are they just replacing it? And I know you’ve done a lot of work around customer complaints. So tell us a little bit about what you’re doing there.

John Sun: Yeah, absolutely. And customer complaints was one of the first use cases that we had deployed. I talked about this on a couple of our own webinars where customer complaints as a use case was really deployed as a design partnership between us and several of the other major fintechs and sponsor banks in the space where we had aggregated data from basically a hundred different fintechs and built tools to help compliance teams manage incoming complaints and better understand what customers are telling them. And I think we really honed in on that as use cases for a couple of different reasons. Number one, it is a uniquely good use of GenAI technology. And GenAI is obviously everything to everyone, as it feels like in this stage of development. But really, what it’s good at is it’s good at processing human language and understanding conversations and compliance use cases, especially around complaints, which tend to have a lot of conversational elements to them. It’s your customer talking to a customer service agent over the phone, email, chat, or text. They tend to be human language conversations, and that’s both kind of what makes them difficult to process. It also gives AI the unique advantage of working within that problem set.

Peter Renton: So, how are they using it? Is the call center or the chatbot, however they’re implementing it, suddenly having 50 complaints come in on the same day? Is that sort of how banks and fintech are looking at it? Are they saying, “Oh, look, this new thing that we just released, there’s some problems here. We’re getting a lot of complaints.” Is that how they’re using it?

John Sun: We’re all about augmenting existing human capabilities with AI. I think the goal, exactly like you said, is how do we strategically insert AI tools and products into existing workflows to make human agents more effective at their jobs? And if we think about the complaint handling process and where some of the inefficiencies or some of the challenges are, it’s really around whether we can train customer service agents to effectively recognize regulatory concerns or regulatory risks in their interactions with the customers. And then from there, can we help human agents better detect when there’s increasing activity in a particular type of complaint? I mean, obviously I think that’s something that humans are already natively doing today. Still, the challenge is for one human agent who doesn’t have access to every single complaint that’s coming in at a given time to identify a pattern and raise an alarm that something might be trending in the wrong direction is extremely difficult. Whereas with the AI tool that can see across the organization to see every single kind of complaint that’s coming in or customer communication that’s coming in, it becomes a lot easier for that tool to be able to create trends of what’s going on and identify the emerging risks or emerging issues.

Peter Renton: So what’s the output from the tool then?

John Sun: Yeah, there’s a few different outputs to this. The first is, again, augmenting that human capability, augmenting that first-line customer service agent or that second line compliance analyst in their day-to-day job. And what that means is using AI to take a customer contact, the customer communication, which is a very unstructured piece of language data and put structure around it in a way that makes it useful for the workflows for that customer service agent or for the compliance analyst. It means, again, taking a customer contact and, first of all, helping the human agent determine if this is actually a complaint or is it a dispute. Or is it negative feedback, or is it an inquiry, or is it kind of just nothing? And that’s really the first step. Then from there, depending on which one it is, we further create more structured data from it. For example, what’s the tone of the conversation, whether the topics or themes we call them identified within kind of that conversation, whether the regulatory threat factors, is there a UDAAP allegation, a fair lending app allegation, a fair credit reporting allegation. Lastly, looking at was this complaint or dispute or negative feedback addressed properly according to the organization’s policies. So that’s really the first set of outputs, if you will, which is structuring the data so that it’s useful to the day-to-day employees handling customer communications and complaints. I think the other output is now that we have all of this cool structured data, how do we help the folks overseeing the function generate better insights about what’s going on with customer feedback and customer complaints. And that’s what we call the insights dashboards. We can show exactly what’s changing within your organization from a complaints perspective. Did you all of a sudden get five complaints about payment refund issues today, whereas you normally don’t get any complaints about that? Does that point to an issue in technology? Does that point to an issue in a change in process or operations and that really allows you to very quickly get to the root of any changes within your product, which might have impact with regards to complaints and helps you identify issues quite frankly, or their magnitude faster than you would be able to rely on existing human processes.

Peter Renton: Really interesting. Yeah. I could see how that would be super useful. These customer service agents, the humans on the front line, are not compliance experts, right? I mean, obviously, there’s training they have to go through. I love the way you framed it: these people, during the conversation, can get a lot more intelligent. They’re not just trying to use their own memory to figure out what regulation applies here. And that sounds really, really powerful.

John Sun: It is a topic that comes up quite frequently as an area that our clients are looking for value from the product. And it’s not just that these folks aren’t trained compliance professionals. They also have 50 other things to do as a part of their job interfacing with customers. So this is kind of one of the 50 things, whereas with an AI trained specifically to do this one task, that’s the only thing it needs to worry about is how to do that one kind of task properly.

Peter Renton: So I want to switch gears a bit. You and I have chatted several times earlier this year about raising the profile of AI-focused solutions in financial services. And you decided to start a conference, which is not an easy undertaking. So why did you decide to start an AI-focused conference?

John Sun: I think I asked myself that same question as we’re T minus 16, 17 days to launch here. And I’m sure, you know, I’m getting a small glimpse into your world for many years. You had to deal with these types of issues, you know, times a hundred, just given the scale of LendIt and Fintech Nexus. But I think, you know, the juice is really worth the squeeze here. I think what we’re trying to do is probably something that’s very necessary in this space because what I’m seeing is that AI is probably the most talked about and least understood topic within financial services today. And, you know, I hear a dozen different ideas every day on how banks we’re talking to or fintechs we’re talking to are planning to implement AI. And the crazy thing is that half of those ideas have nothing to do with GenAI. It’s traditional machine learning, or it’s basic analytics. And the other half is basically all about replacing call center agents. And there’s just so much more depth to GenAI and then the non-GenAI stuff and then purely replacing call center agents. And we thought it was the right time to open up a discussion to have folks share what they’ve been working on. That’s a little bit outside of the box. What’s working, what’s not working, and disseminating real, practical information and practical implementations of hopefully imaginative use cases. And I think the other side of it is the regulators; I think we need to have a conversation with the regulators because there’s a lot of regulatory activity in the area. We’re at a moment where the regulators themselves are still trying to figure out what they want to do with the technology and what they want us to do with it. So, as a result, there’s this narrow window to really share and promote what some of the more consumer-friendly use cases of AI are to help shape their understanding of what’s possible and, in turn, to hear from them where their focus is which helps us shape our focus on what we’re trying to do with the technology.

Peter Renton: So you decided not to locate this in the obvious places like San Francisco, New York, or DC. You decided to put it in Utah, not the first choice I think that most people would think of when it comes to a fintech banking type event. So why Utah?

John Sun: It’s a great question. And I think a lot of folks in the fintech space will  look at Utah, and like you, immediately go like, Utah? There is a little bit of background here. Utah has been at the epicenter of fintechs for quite a while. It’s just not a lot of people think of it that way or know about it that way. A lot of the sponsor banks that power most of the brand name fintechs, lenders and neobanks and credit card companies are based in Utah. It’s a combination of a few different aspects, obviously, favorable regulatory environments. But also really just a conscious effort by the Utah government to prioritize fintechs, which makes it an ideal home for the growth of fintechs. And really the reason we’re basing it in Utah is I think it has great sponsor banking presence. I think it has great banking presence. A lot of the fintechs that we want to share their experiences on AI adoption probably make their way out to Utah to meet with their sponsor banks on a pretty regular basis. So it becomes a pretty natural location to have the right folks represented all in one place.

Peter Renton: And it seems like the governor’s really gung-ho here because I saw the Utah governor is speaking at Money 2020 later in October. So that sounds like from the very top, Utah is really focused on this.

John Sun: Yeah, absolutely. And that’s kind of what we’re seeing as we were planning this event as well. We’ve gotten just an incredible amount of support from the folks within Utah. Specifically, to call out the Utah Governor’s Office for Economic Development, the Utah Bankers Association, the Stena Center at the university there. I think they’ve done a great job of aligning a lot of different organizations within Utah behind this goal of making Utah the epicenter of fintech and banking growth.

Peter Renton: Okay. So, who are you trying to attract? Who is attending this event?

John Sun: I think the main audience we’re trying to attract is the fintechs that are building cool, cutting-edge technologies for consumer products within finance. The sponsor banks backing these fintechs as well as other banks that are looking into AI adoption, as well as the regulators thinking about how to use AI to better enforce their policies or how AI could be harmful or beneficial to consumers. And really, it’s also the three parties to the conversation that we want to have. That being said, I think the event is designed to be very practical, not high-level theoretical about AI. So within these organizations, really, we’re looking for practical leaders like folks in executive leadership or compliance or operations or marketing, folks who are going to benefit from the technology and need to understand it from a ground level. And I think that’s where the main focus is. But over time there’s been a lot of other kinds of interested parties and interested groups that’s become a part of the journey here.

Peter Renton: Okay. So then, let’s talk about the agenda. What topics you’re actually going to be covering. Maybe just run through some of the highlights there.

John Sun: Yeah, absolutely. I think the focus again is going to be on practical application. So we’re going to be talking about things like what are the AI tools or options available for using AI within compliance or using AI within customer experience or customer service or operations. We’re going to be talking about the practicals of how do you implement AI within the organization and what are some of the compliance considerations in implementing AI from a model risk governance or a safety and soundness perspective. And I think we’re going to hear from the regulators on their view on AI adoption and the types of things they’re going to be looking for as a playbook develops for AI adoption within financial institutions. We’re hoping a lot of the conversation is going to come from our excellent set of panelists, moderators, and keynote speakers as well as from the audience who are going to inform whether the types of topics that are going to be relevant to the development of AI adoption within the coming months and years.

Peter Renton: I’m looking at the agenda on my computer here. You’ve got some heavy hitters. You’ve got Brian Brooks, the former acting Comptroller of the Currency. I think he is one of the smartest people in all of financial services. I just love listening to him talk. And then you’ve also got the vice chairman of the FDIC, Travis Hill, looks like he’s going to be interviewed by Phil Goldfeder, our friend from the American Fintech Council. So you’ve been able to track some decent names here, and I see you’re going to be featured as well. John, when you’re looking at the agenda and the day in general, are there things you’re most excited about personally?

John Sun: I am, by experience and most of my career, a huge data geek, so naturally, I’m going to be drawn towards the more data-centric topics. I think there will be a lot of practical data about implementations that have happened as a part of much of the agenda. I’m pretty excited to hear from folks on what’s actually been going on, the numbers behind the stories behind AI. Obviously, we’ve seen some of the early headline numbers that folks in the industry have been releasing about the impact of AI within the industry, but we’re really excited to see what direction that’s headed within the specific organizations of the participants.

Peter Renton: I’ll link to the event in the show notes. Are tickets going to be available right up until the last minute?

John Sun: Yes, tickets are available right up until the last minute. I definitely strongly encourage everyone to grab their tickets if they haven’t already. The prices do go up at the last minute. But even if you haven’t had a chance to purchase a ticket yet, there’s definitely still time. Please do that. Would love to see you guys in person in Salt Lake. I think it’s going to be a great event. I will also call out that we set aside a limited number of free tickets for industry experts and practitioners in the spaces we want to highlight. And some of those are still available. So, if you’re interested, I strongly encourage you to go on the website, it’s conference.springlabs.com, and apply for one of those free tickets and spots available.

Peter Renton: Is your plan to have this become an annual event, or is this just a one-off?

John Sun: You know, we want to have the conference for as long as the topic continues to be relevant and interesting. And I think AI tech will be at the forefront of innovation for many years to come. I think the format of what we talk about is probably going to change from year to year based on what the latest needs within the industry for AI adoption, the understanding of AI is. But I think it’s safe to say that the conference will be making a return even if the topics will be different.

Peter Renton: You’ve got this event, which will help drive the industry forward. In closing, I’d love to get your vision for Spring Labs. Where do you hope to take this?

John Sun: Great question. It’s a question I ask myself oftentimes, and I’m not sure I have the best answer, but I will tell you a bit of the background behind what the mission we’re trying to achieve at Spring Labs is. If I think back to what we did well at Avant in the early days, this was again, the very early days of fintech in 2012, 2013 when it was basically just us and Lending Club and Prosper. And we made up essentially the prime and near prime fintech space. What we did really well was lean into the latest cutting-edge technologies. At the time, it was machine learning technologies like gradient-boosting models and what’s called XGB models. And that gave us a source of competitive advantage, but it wasn’t without cost and peril. There was no infrastructure for deploying this type of technology. So we actually had to build everything from the ground up. And it’s an investment that most companies wouldn’t have been able to make. Certainly, most banks wouldn’t have known how to make. Fast forward to 2018 when I started Spring Labs, the idea was, how do we, again, make it easy to adopt the next generation of these technologies a form that is easy to implement, as well as ROI positive from day one. And that’s really what we’re hoping to accomplish with a lot of the AI products that we’re building to create an easy, attractive, revenue-generating, or at least ROI-positive way for financial institutions to adopt this great technology in generative AI that will likely be transformative for many years to come.

Peter Renton: We’ll have to leave it there. Good luck with that, and good luck with the event on October 7th. I’m very disappointed I actually can’t come myself because it happens to be my daughter’s 16th birthday. Otherwise, I would definitely have been there, but best of luck with everything. I hope it goes well and thanks for coming on the show, John.

John Sun: Thank you so much.

Peter Renton: I hope you enjoyed the show. Thank you so much for listening. Please go ahead and give the show a review on the podcast platform of your choice, and go tell your friends and colleagues about it. Anyway, on that note, I will sign off. I very much appreciate you listening and I’ll catch you next time. Bye.