Saving Half a Day with Trovata TMS: A Chat with OneMain Financial

This transcript was auto-generated. We made some light edits for clarity and readability, however it may still contain some errors.

Joseph Drambarean 

All right, welcome again to Fintech Corner. I’m joined today by Steven DeLutri, CTP from OneMain Financial.

Thank you so much for joining us here on Fintech Corner. We’re here at AFP. And I was hoping that we could kind of just jump into it, talking a little bit about your experience kind of bringing technology to life in your organization. 

You’ve been at OneMain Financial for 10 years, multiple acquisitions, a lean team, self-described. How has tech played a role in helping you do your job?

Steven Delutri

First off, thanks for thanks for having me on the podcast. The way we think about it, at OneMain, it all starts with the people. 

We do operate a lean treasury team, and when you operate lean, you do have to have good people. And we always put that ahead of the technology. Once we have good people, we know what to do with the technology. 

So we’ve developed a process when, we run through checklists. We engage our internal teams early, our banking partners, our treasury management system partners like ATOM and now Trovata TMS. And we really just put a lot of diligence in on the front end because when you work with the systems, like anything else, garbage in, garbage out. So we have to be really diligent about getting all of that done up front.

Joseph Drambarean

Absolutely. All of the stories that have come through this booth have been really around that, that synergy, just like you said, between good people, but also good technology that kind of brings that bridge together so that you gain efficiency, you know, and automation.

How has some of that played out in your use of technology, especially when it comes to data, when it comes to managing risk, payments, of course another big area. How has that played a role?

Steven Delutri

Yeah, it’s been huge. Data is very important. The real time nature of APIs plays into that as well. We’re able to be very efficient in how we process our payments. We get responses back in near real time so that we know that transactions are complete. 

Cash positioning is like seamless, basically done when you walk in the door every morning.

You know, we started 11 years ago with Excel spreadsheets and 15 banking partners and it was a half day just to figure out how much money we had and now we just walk in the door every morning and say, hey, you know, here we go.

Joseph Drambarean

It’s just magic. It’s just there. What does that do for the business ultimately? Because that transition from having to do that as a half day activity to it just being done, does that allow the business to grow at a faster pace? Does it allow new business opportunities?

Steven Delutri

Well, you for our team within the treasury organization, what it’s what it’s really done, it’s allowed for the people to evolve and, you know, work on new skill sets. Now we’re into more value added activities. We’re no longer we’re no longer, you know, making the doughnuts, so to speak, where we’re doing the more value added things, the more analytical things. And then we are able to work more cross functionally across the organization as a consultant and and really help the business.

They’re looking to grow the business, streamline the business, look for cost savings. So not just in the treasury area, but across a broader landscape of the enterprise.

Joseph Drambarean

One of the things that has influenced a lot of the evolution in the space has been the adoption of APIs, as you mentioned, and getting nearer to that infamous “real time.” We’re inching ever closer, ever closer to that possibility. 

Have you seen any challenges in adopting APIs, and how has that journey been for you?

Steven Delutri

I think the biggest challenge we’ve had at OneMain is that the banks are all moving at a different pace with different priorities. And we’ve recognized the value of APIs early on and the banks where we have done the APIs with the ATOM system and, you know, what is I guess now Trovata TMS. It just works so much better.

When you have those time-sensitive payments and you’re able to confirm and provide details to legal when it’s an M&A or some other sensitive type of transaction, it’s just great. So we want to do more of it. So we love the speed, the efficiency, we’re not scared to jump in on new technology. We do a lot of testing, of course. So yeah, we love it. We see real value, know, time is money.

Joseph Drambarean

One of the things that has been fascinating in the evolution, especially over the last five to 10 years in this space is that where typically in the past information reporting or payments or all these different aspects, they may have been important and adjacent to the primary banking relationship, but it was never kind of the main thing that was discussed. was secondary to the rates, the arrangements that you have from a credit perspective, etcetera. Now it almost feels like it’s starting to shift where the conversation around technology; APIs, the availability of them and those services, is becoming a priority that is at least equal to some of the rates and services that you will receive. Have you seen it as kind of a driver of the choice that you will make in an RFP with a particular financial institution?

Steven Delutri

Yes, absolutely. When we’re looking for a new service, one of the first few questions we ask is about how has their technology evolved?

It’s become a real differentiator in banking services. So yes, we look to give business to our banks who provide us credit. That’s a big part of our company and the industry we operate in. But we can’t just give business to a bank if they don’t have the services that we need. The APIs, how the bank is thinking about AI, all factors in.

Moving up the list I think every year and exponentially. So we again we think it’s front and center to what we do.

Joseph Drambarean

Now that you’ve had experience with APIs for it sounds like quite a bit of time and they’ve driven leverage, right? Speed being the primary one. Have there been any areas where you see kind of hopeful development and evolution of the technology? If you were to have your top three requests fulfilled, what would they be in the space of APIs and maybe just generally in the area of data?

Steven Delutri

I think for starters, it would be great that we would have some parity across all banks so that we have a little more ease of implementation. So we rely on companies like Trovata so that we don’t have to manage all the differences from bank to bank. So definitely parity across banks would be great. And then, access to more data; Thinking about like OFAC, KYC, the onboarding process when you open a new bank account… that is something that a lot of companies struggle with with their banking partners. It’s a very slow process because of the KYC. If there’s any way we can leverage APIs to speed up that on the front end, that would be game changing stuff, I think, for a lot of companies.

Joseph Drambarean

Digital onboarding seems to be a priority in every area of the finance industry except for in corporate. And it feels like if there was an evolution to happen, that would be, as you say, game-changing.

Steven DeLutri 

OneMain has invested a lot of money in creating that experience for our customers, right? As a consumer lender, but yes, back in the back office and corporate functions were a step or two behind and we would love to get there.

Joseph Drambarean

One of the things that is very predominant on the showroom floor here is the conversation around AI. And it’s easy to generalize, you know, well, AI is going to solve everything. 

But I was curious to hear your take on AI and the role that it can play in kind of your day to day. Have there been any areas already that you have been experiencing the benefits of AI? If not, where do you hope to see the benefits?

Steven Delutri

We’re definitely stepping into it lightly. We, as a company, we are all in, but we’re moving cautiously to make sure we understand and know how to use it, how to prompt it. Just like any other technology, garbage in, garbage out. So I think that’s an important thing to understand. That how you build a prompt is gonna lead to the results you get.

We see AI as something game-changing and if we implement the right solutions with the right partners, we see that we can use it, you know from helping with loan underwriting to iImproving fraud control, whether it’s corporate payments or the consumer payments. So there’s a number of different areas where we seek to embed AI. I think fraud control is a big one, regardless of where we sit in the enterprise.

Joseph Drambarean

Have you been exposed to any conversations internally with regards to data governance, especially since you already kind of are sitting in a position of managing a lot of data from the banks and being responsible for that data being made available anywhere else in your business systems. As AI becomes a powerful topic, how has that played into kind of data governance and how data will flow throughout systems?

Steven Delutri

Yeah, I think data governance is very important. Like we’ve always, you know, where we sit in treasury and we always, whenever we built out a treasury management system, the idea of like having your master data right is very important. So I apply the same kind of mindset as we like shift into like AI and data. We have an internal team that is going to provide that framework for the whole organization no matter where you sit in the organization so that there’s a common set of rules that the company will play by in the AI space. 

Every vendor, whether they’re an explicit AI vendor or a vendor that is introducing AI into their product, will have to go through some vetting process. So we think that’s important and everyone should be thinking about that. You can’t just jump in and start using it. There has to be some controls around it.

Joseph Drambarean

One of the things that I found challenging as a conversation very recently, and just sharing a little bit of the inside baseball of our day-to-day operations was a challenge from a different customer around the concept of AI and the role that their treasury team will play in governing accessibility to their treasury data to a model context protocol server, an MCP server that would be used by another large language model and the challenge to us was if the Trovata data infrastructure is our data governance platform, how will it then lend access to all of these large language models? 

If we want to play with ChatGPT over here in one organization and Gemini over here in another, and maybe one more, Claude, over here, how are we going to maintain consistency when feeding data into all of those different systems?

It definitely put me in a position of sitting back in my chair and thinking, wow, that’s a problem that is probably going to be very similar across organizations. It’s going to be real over the next five years, because I’m sure every organization wants to use these tools. And the last thing that, that you want is each of them creating their own fiefdoms of data warehouses that feed their own AI programs that may or may not have consistency across the organization. 

I’m curious, has that kind of conversation played out at all at OneMain?

Steven Delutri

Yeah, something that we realized early on is you can literally write the same prompt into five different tools and get five different answers. So there needs to be some testing around that. We’re using the co-pilot. I think a lot of large enterprises are using co-pilot. We don’t have all the bells and whistles turned on.

We’re stepping into it lightly, but I wonder from where Trovata sits and where a company like OneMain sits, we like to ingest data into our data warehouse and just let us then do what we do. So we’ve not really done that with the treasury data, but I can see that happening. 

We use Snowflake and we can just ingest data and just use tools like Power BI and have models running in between that could be a thing, know, just just thinking out loud.

Recommended: Snowflake’s Treasury Team Drives Speed and Accuracy with Trovata’s Cash Reporting 

Joseph Drambarean

No, it’s actually very prescient as a thought because data portability is becoming a very common theme for our large enterprise customers. For the mid market that maybe they’re just stepping into treasury and they’re not yet at the point of thinking about that kind of distribution of data, it’s less of a theme. But for large enterprises that have multiple finance programs that have their own business systems that need this kind of leverage, it has come up and Snowflake, Power BI, obviously another kind of platform in Azure has come up as, know, if we already have warehousing systems and analytics systems and AI systems, how exactly are we going to get a carbon copy of some other database that might be, for example, in Trovata TMS and have that serve our needs everywhere else? 

It’s an interesting conversation because in the past, the way that we’ve approached it has been, well, you can access that data any time through APIs, right? Just go get the data from the system and use programmatic interfaces to do that that are well governed and that are well entitled. 

Well, it’s transitioning now to, well, we just need that data in our own warehouse because we don’t want the risk of bringing down that software stack because some AI model wants to run rampant and, you know, look for things. So it’s put us in a position of really rethinking the architecture and making it available in a more, you know, accessible format for these types of systems.

Steven DeLutri

You have to have multiple flavors of that, like you said, because of the different size of the company and what their capabilities are. So if you want to make your product appealing to companies of all sizes, you then have to have tier one, tier two, tier three, and maybe some kind of hybrid model as well.

Joseph Drambarean

Exactly. It actually makes me think of another question. This is fun because I feel like we are riffing off of each other. The position of data governance isn’t one that is written in the playbook of treasury, right?

But, seemingly, it’s starting to become an important role because if financial data is to be trusted, someone needs to be responsible for the veracity of that data. And if it’s being stored in a technology, whether it’s a data warehouse or a TMS or whatever it might be, someone is responsible for making sure that that system is up to a certain level of, you know, reliability. Has that changed your role?

Steven Delutri

I don’t know if it’s changed my role. I think we’ve always taken the approach, taken a very serious approach around governance. We have a very rigid third party risk management process. All key vendors have to do like an annual recertification, both at a general risk level and a cyber security level. So the data governance… I think within Treasury, I think we’ve built the right rules on it. Because if you go back before we had all this technology, there was a lot of, you know, audit requirements that govern that you have to keep certain sets of data for seven years or 10 years. 

So like, we always had that infrastructure. So now we’ve just kind of built that into the cybersecurity landscape as we digitized everything. And, and, you know, those rules exist throughout and

You know, so I don’t I don’t I don’t think it’s really changed for us, but I could see like smaller companies and mid-sized companies like being a little You know, as you know as they move through, know implementing newer technologies, you’ve got to really think about the framework, the governance, and not just let it run wild.

Joseph Drambarean 

It’s interesting because you are so fluent in what are complex IT topics, data warehousing, data governance, as we’ve been talking about the use of that through APIs, connecting these systems. That experience, do you feel like your background at OneMain has gotten you to the point of having that experience? And do you think that that is going to become standard practice for the treasury role moving forward?

Steven Delutri

I think so. I think for any functional area within an organization you have to have some level of expertise in technology. So, I’m not a technologist by any means, but we do work cross functionally across the organization every day and we have key partners that work explicitly with our team from technology and we’re able to bounce ideas off each other. 

We’ve evolved as a company in that manner and we have people on our team who are versed in the data warehouse and how to write scripts and access data in Snowflake so we don’t need, you we don’t have a heavy reliance on analytics themes. They’re focusing on our core business. So we’ve built that expertise within treasury. I think that’s a real advantage for my team. And I think of things should, you know, maybe do the same.

Joseph Drambarean

Speaking of building that advantage, would you mind if we play a little bit of futurist and look at maybe 10 years from now? Let’s say that we believe all the hype, okay? And agentic systems exist and they’re reliable.

And we’re in a world where we have Jarvis from Iron Man, and we can speak with an agentic system as we do today, but maybe more importantly, we can trust an agentic system to be doing work without us, right? Given a general directive, let’s say that this agentic system is on your team. It’s playing an associate treasury role and it’s been given a job.

What kind of leverage do you think would exist if we could get into that future? And do you see that as even a possibility?

Recommended: Why Agentic AI Isn’t Hype: Real Treasury Tasks It Will Soon Automate 

Steven Delutri

I think maybe 20, 25% of my thoughts lead us in that direction. Others might be thinking more aggressively. I still think a little old school and, you know, we work in treasury, we control the company’s money and we want to make sure everything is on the level.

I think there’s a lot of work to be done before we can get there. We are we are exploring agentic tools for our consumer facing stuff. I think a lot of lenders are doing that. But in the back office and corporate treasury, I think we’re a little ways off from from getting there, although, you know, like I said, 25 % of me thinks we get there at some point and

I think if we were to get there, I will still rely on people who have the functional expertise and just will use the technology to make us even more efficient so that we can just do more and be more consultative across the organization. I think that’s the mindset we’ll have.

Joseph Drambarean

I’ve really enjoyed our conversation and it honestly makes me feel like you have so much wisdom that you could share to our audience. If I were to ask you to kind of give your hottest take your best piece of advice for anyone that’s kind of walking the showroom floor, trying to figure out what to do next. Maybe they are in a position where they have to make decisions around technology and a new vendor and all of that. What advice would you give them?

Steven Delutri

The advice I would give anyone is have the conversations, have more than one, multiple partners and possibilities. Do your homework and do your due diligence before you jump into a solution and don’t be afraid to ask questions. No question is a stupid one. I don’t know how many times we’ve all heard that in our life, but it’s so true because once you start moving into a technology solution and building it out, some of the very first steps you take will set the tone on where you go with it. And you want to get those first steps right. So just make sure you do your homework and diligence.

Joseph Drambarean

Well, Stephen, this has been awesome. Thank you so much for your time, and we appreciate you.

Steven Delutri

I’m happy to do it anytime.