Ep. 13: How is AI in banking changing risk and compliance? Rens Troost of Rational Exponent explains

[music]

Grant Evans: Welcome back to The Payments Shed Podcast. Today we are joined by Rens Troost from Rational Exponent. Rens, please introduce yourself.

Rens Troost: I’m Rens Troost, CTO and Founder of Rational Exponent. Our job is to provide prudence as a platform and to bring you the bank of the future today, and we can talk about what that means.

Grant Evans: Everyone’s talking about AI in particular. Where are you guys seeing results with AI in your world?

Rens Troost: A lot of people are using AI for fairly shallow use cases. People are using it for content generation. People are using it for vetting marketing messages.

We’ve gone to a more interesting place. What we’re trying to do is look at broken bureaucratic processes, which are the gating systems at the heart of every bank, and make them a whole lot better.

The place where we’re starting is in second-line-of-defence risk management, which sounds like a mouthful, but it actually is what stops progress from happening at most organisations. Yet it is a critical function that must be performed.

Grant Evans: Trust is obviously a big part of what we talk about in payments and the regulation side of things that sits around that trust point. How do you guys look at the compliance side of things with your platform?

Rens Troost: Our platform automates the processes for both prudence and compliance in banks, so it’s actually right at the heart of what we do.

If you look at some of the things that are going on in the industry right now, for instance Revolut’s ongoing struggles to get a banking licence in the UK, you’ll realise that as a fintech, or even as a traditional financial firm, there’s nothing to be gained by trying to go around the system or short-circuit it.

You have to go right at it. Going right at it means complying with the rules as they are, but doing it in a new way. We bring that new way, and that’s what we’re doing.

Grant Evans: You talked about AI being fairly shallow in some ways. Where are you seeing good use cases of AI being properly implemented and used for a lot more beneficial processes?

Rens Troost: All of the software developers, including the ones on my team, wonder how they did their job before AI. Absolutely, that’s the case.

But what we’re doing with our platform, Reagent, is basically infusing the risk operating model of a bank with AI and turning it into a closed loop where you can make sure that your policies conform with regulations, make sure your actual practices conform to the policies that you believe you conform to, and evidence that to the regulators when they come to do an examination.

What I find most exciting is, in a closed loop, being able to affect the product introduction and product approval process.

By using AI in the right ways, and we can talk a little bit about what those right ways are and how they preserve accountability across all of those processes, what we hope to be able to do is take all of the smaller banks, all of the non-traditional banks, all of the people who are disadvantaged by the current manpower-heavy way of doing this thing, and have them be able to punch with the big boys.

Grant Evans: Do you see a difference between legacy banking providers and the more neo bank?

Rens Troost: I think yes.

If you look at the big five banks in the United States, they have 10,000 people plus in their risk functions.

If you look at smaller banks, fintechs, a bunch of the new DeFi providers and all that kind of thing, they can barely spell compliance as it’s spelled by the regulators.

There’s a very big difference in culture, history and understanding. Yet over time, whether it’s a deregulatory administration or a stricter administration, you need to be able to play the rules of the game as they are set.

We hope to be able to level the playing field for all of these people and do so in a way that’s just a fantastic place to work, which is not always the case in risk these days.

Grant Evans: I get the feeling you’ve maybe got a few opinions on some of these providers in the market, but are they not also addressable customers for you?

If they operate in a way that you deem to be slightly non-compliant, or they’re struggling around that compliance piece, they’ve got to be a great customer fit for you guys, right?

Rens Troost: Absolutely.

There’s a whole legacy ecosystem of products called GRC systems, and there are a lot of workflow products that are used in the management of risks, the management of exceptions, the management of transaction flows and new product introduction.

But a lot of them are wrapped around and built for an operating model that is incredibly human-heavy.

One of my touchstones is an old book from the 1960s called The Mythical Man-Month. Its observation was that the productivity of a team plummets geometrically as the size of the team expands.

What do you think that says about a 10,000-person risk management team?

AI is not going to replace everybody’s jobs. But what it is going to do is take the best people in your organisation and amplify them to where they can do the jobs that would otherwise require 10 men or 10 women to be able to do.

In the future, we’ll be able to have a much more streamlined organisation and the AI itself will maintain the model.

This is critical about trust. AI will maintain the model of accountability that the regulators are comfortable with.

Nobody wants, and in European law one is prevented from having, unaccountable computer systems making decisions that can’t be challenged.

We maintain the human accountability by amplifying the humans. The actual people who are responsible for independent risk management, compliance and prudential activities in the bank can have their own decision process mirrored by the agents that work in the multi-agentic framework that Reagent brings.

That’s a mouthful, but what it means is technology that is under the control of people and makes people more productive and happier.

Grant Evans: I think it’s a great way of putting it. There’s always been this innate fear that AI is going to replace everyone and everything. But the way you refer to it, that it will make your best people better, is a great way of looking at it.

Rens Troost: Across human history, every technology introduction, from the introduction of the looms that the Luddites fought against, and probably to the introduction of the wheel that we don’t have anything about, has had winners and losers.

But overall, it’s created more opportunities than it has destroyed. I think that’s going to be the case with AI as well.

What I would like to see in the future, and I’m a very optimistic person, is that we’re going to have a world where a lot of people are empowered to do things that formerly they would have to become a small cog in a very large machine to affect any part of.

More of that, please. That’s what we’re trying to bring right now with Rational Exponent.

Grant Evans: We’re obviously here at Money20/20 in the USA. Why is this such an important event in your calendar?

Rens Troost: In my calendar, this is possibly the most important event of the year, aside from the meetings that we have with our customers.

We’ve been working for about two years on Rational Exponent, for about 13 months on the development of the platform that we are launching here at Money20/20.

Reagent is now open for business as something that anybody can buy. We have already deployed it a number of times in banks, so we’ve had our early adopter customers.

It’s fantastic to be here with basically the entire innovation ecosystem for the financial services industry.

A lot of the people who come here come for things that are obviously exciting innovations. They come here for crypto. They come here for stablecoins. Way too many people come here to sell identical products around fraud and payments and those kinds of things.

Many of those products are fantastic, and many of them are going to sink without a trace.

But it’s the people who are here, not at the booths and not on the stages in general, who are in the banks looking for an advantage, looking for how the system is going to work going forward.

Especially right now, in a time of massive regulatory change. Regulatory change here in the US, of course, regulatory change back home in England. There are all kinds of things happening.

What Reagent is aimed at doing is, whatever direction the regulators steer us in, whatever direction the innovation ecosystem ends up making affordances that drive you one way or the other, we help you to be able to adapt to it quicker.

So, bring on the change.

Grant Evans: How do you find dealing with completely different regulations in different markets?

Rens Troost: It’s challenging.

As a result, I do what I frequently have done, which is introduce products in the United States because it’s bigger and because it’s actually a better place to take risks.

I actually love the way that Britain manages innovation, and I really love the way that the PRA works and the Bank of England works to foster innovation.

But look at the size of the market here. Look at the size of the ecosystem.

We’re striking partnerships with a lot of fintechs and other non-traditional banking organisations here as well. You’ve got to be in the United States.

Grant Evans: That’s a fair point. I think we see with the UK a lot of companies start in the UK and then come to the States. But doing it this way, you’re maybe getting over the hardest part first, right? Then you roll it back into other markets from the US.

Rens Troost: Yeah.

It’s also the case that the AI ecosystem is here. There’s some fantastic talent here. There are some fantastic vendors and partners.

It’s a moving target what the regulations around AI are, but we know what the risk frameworks are.

By the way, the American standards organisation, NIST, is fantastic at laying out how to think about AI risks. We’ve imbibed all of that in the creation of our systems.

Because if you’re going to build systems that help banks manage risk, be prudent and attain compliance with a bunch of things, you better do it in a way that doesn’t introduce new risks that you don’t understand.

Again, the Americans are leading a bit in this area.

Grant Evans: Fantastic. Just before we finish up, there’s a little segment we do at the end of every chat that we have with our guests called The Shelf of Shame, where you nominate something that really grinds your gears, whether it be in payments, business, fintech, whatever you want to choose.

What’s going to be your nomination for The Shelf of Shame?

Rens Troost: I think a lot of people probably are going to share this one.

Every company now, if you look at their television ads, if you look at their marketing materials, they’re all AI companies, aren’t they?

Grant Evans: Funny that.

Rens Troost: I like real AI companies, companies that were born to actually take advantage of the new AI that’s been around since 2016.

People who say they’ve been doing AI for the last 25 years, sure, a lot of banks have been doing machine learning for the last 25 years. All banks, you might say, that were around 25 years ago have been doing machine learning for the last 25 years.

It’s an entirely new ball game.

So it really grinds my gears when everybody tries to AI-wash themselves.

Grant Evans: Fantastic. Thank you for coming on The Payments Shed.

Rens Troost: Thank you so much.

Grant Evans: See you soon.