Don Bowen on AI’s Impact on FinTech

In this episode of Lehigh University’s College of Business ilLUminate podcast, we are talking with Don Bowen about the impact that artificial intelligence (AI) has already had on the fintech industry and some of the ways that AI offers both future potential benefits and threats to fintech.

Bowen is an assistant professor in the Perella Department of Finance who teaches Lehigh Business's Capstone course in FinTech, among other courses. He also is co-director of Lehigh's FinTech Minor.

Bowen spoke with Jack Croft, host of the ilLUminate podcast. Listen to the podcast here and subscribe and download Lehigh Business on Apple Podcasts or wherever you get your podcasts.

Below is an edited excerpt from that conversation. Read the complete podcast transcript [PDF]. 

Jack Croft: As I mentioned in the introduction, the impact that AI has already had on the fintech industry, let's start there. What kind of impact has it already had that you've seen?

Don Bowen: Tremendous. It's been pretty profound already, and it's only going to accelerate from here. AI, I think of it as having two main components. One is a set of tools that make it easier for firms to do predictive analytics. So they're trying to understand some outcome variable, like will this customer pay back their loan? So that's one facet, predictive analytics.

The other facet is generative AI. So these large language models (LMMs), such as ChatGPT, and integrating these language models into how the firms do their many different tasks across the firm. On the latter - on using generative AI tools - it's going to change not just the way employees at firms do their job, but it's going to change the way customers interact with firms in tremendous ways.

The first one thing that's already underway is the broad adoption of chatbots for customer service. So Klarna is a loan provider, and they recently disclosed that for about a month, they've been using an AI assistant for customer service chats.

And in one month, it handled two-thirds of their customer service chats. Basically, it was about 700 employees’ worth of work. And humans, the customers, gave it nearly equivalent to satisfaction scores. It did a great job solving customer inquiries. A 25% drop in repeat inquiries. And it was solving customer problems much faster. So I think the number they gave is it was about 80% faster at fixing customer problems. Because it's a computer, it's on 24-7 around the world, in a lot of languages.

And Klarna thought, OK, this one use of AI - one use - is going to help them increase their profit by $40 million because obviously they're going to reduce their customer service organization. So the wage bill will fall. But also, you can bake in some value from the quicker service and the happier service that they get. And that's just one way that these GenAIs are going to be useful.

Croft: So if you could talk a little about that balance between a more efficient, faster function with AI and the impact it has on the human workforce.

Bowen: The first thing that I would think about in considering how AI is going to impact workers and people more broadly through their other relations with the firm production, it's going to make firms more productive. They're going to produce more goods at a lower cost, and consumers are going to benefit from that.

This is true if I'm talking about AI now, but it's also true if I'm talking about automated assembly lines or talking about using cars instead of horses, right? So new technologies invented, this changes the way things are made, and the way things are done. And the first and largest impact is this makes everybody better off. We are able to make more things with less. So the pie is a lot bigger now.

There's always some questions about who is affected by that, and how long are they affected, and how deep is the effect. This changes from generation to generation as we have new technology that comes online maybe has slightly different fundamental attributes than prior ones.

But it also depends on the macro economy. Right now, if you get displaced by generative AI, you're going to be on the job market at one of the best times in history to find a job. Our unemployment rate is historically low. And job seekers are having an easier time than ever at finding new employment. So that at least makes it likely that the impacts are going to be less severe than they might otherwise be. But it's a very interesting topic to think about.

Croft: Let's talk then about some of the ways that you see so far that AI holds the promise to improve the fintech industry. What kind of things that it may not be capable of doing today, but the trends are that in a short time it will be able to do?

Bowen: So just a few promising avenues that AI could be used to influence the industry. One, enhancing customer experiences, which we just talked about one example of. Two is refining risk management processes at firms. And the third one is hopefully improving financial inclusion.

One of the first things I said about AI is that one of the things that AI means to me is predictive analytics to try to understand some outcome, like is a borrower going to pay back their loan. Financial firms have used AI techniques for quite a while to improve their risk management processes. So they have a lot of data. It helps them understand which customers want which products and which services. And it helps them improve their pricing models.

And the nice thing about that is that, in a sense, two things happen. One is that more people are offered credit. So some people who previously weren't offered credit are going to be offered credit. It might be at a higher rate, but their access to credit will expand. And that can make the lending space more competitive, which is better for customers.

And the access to finance, we know from a long history of research, the access to finance is very, very important for the dynamism of an economy and helping new firms get founded and grow, helping people manage their personal liquidity so that they can concentrate on other things that matter to them, right?

So if you're worried about making your next bill and you're having trouble raising funds, this burden is going to make it harder for you to concentrate on your family and your children. Finance has always been a backbone of the economy, and things that improve it are generally going to have tremendously positive first-order effects.

In the finance space, we get risk management techniques that are helping firms assess loan applicants. And also, in the first wave of AI implementation in the finance industry, fraud protection expanded a lot. My first experience with this was back in 2008 when Wells Fargo told me somebody was trying to use my credit card in Lafayette. I had never been to Lafayette. And they saved me that day $300. And I was an undergraduate student at the time, and that felt like saving $1 million. [laughter]

Nowadays, these fraud detection mechanisms are much more sophisticated. When you leave the country, you don't need to tell your bank that you're leaving the country. You used to have to tell them, and now you don't. And when you go abroad, your credit card's going to work. But if somebody in a different country were to use my credit card right now, it's more than likely that my bank would flag it for me with relatively fast speed and protect me. So that's the risk management.

I guess the last thing that I mentioned at the top was the idea that AI can improve financial inclusion. So one thing I said earlier was that by allowing firms to do a better job of understanding the risks for borrowers, that makes it easier for them to open access to more borrowers because they can price their loans, that credit risk, appropriately.

The other thing that's really exciting is the expansion of robo-advisors and hybrid advisors. A lot more Americans than in the past have access to retirement funds and savings. But making choices for your retirement accounts is a very complicated thing for most people. And the people who study this find that a lot of people leave a lot of money on the table by making some relatively simple, easy-to-correct errors, to be blunt, with how they set up their retirement accounts.

There's been a large growth in the use of robo-advisors, or nowadays, hybrid advising systems. So it's an AI system that's suggesting that, "Hey, Jack Croft, he should probably consider reallocating his money from one investment type to another." But what happens at several fintech firms or firms that are adopting this technology is that actual advice is sent to a human advisor to tell the customer. So there's a human in the loop.

The generative AI and AI system that they have kind of analyzing your retirement account and the way you're set up will generate some insights, some suggestions. Those will be sent to a human who can review them quicker because the human doesn't have to come up with all the ideas from scratch, but simply sits there as a reviewer and then tells that to customers, which customers love.

People like advice from LLMs, but they like it less when they know it's from a robot. They like advice from robots more than advice from humans. But once you reveal that the advice is from a robot, people get a little bit hesitant. So what finance firms are doing these days is changing the way it works to often have a human in the loop.

For example, at Vanguard, if you're in one of these systems, the advisors who are doing this kind of hybrid advising, they don't service the same number of customers they used to service. They service a lot more because the AI system makes it easier to do their job.

If they want to work with one customer, it's a lot quicker. And so they work with more customers, which again expands access to that advisory service that did not used to be accessible before. There used to be a very high net worth minimum to get into those programs. And now it's lower because they can do their job more cheaply. So that expands access.

And not only did that expand access for customers, that actually expanded the amount of employees that the firm needed. Because now, instead of only servicing some elite customers above some high net worth, they're now able to service customers at a much lower net worth. And there's many more of those customers, which then requires more advisors.

So this kind of gets back to what we were talking about earlier, the way AI people are worried that it might displace workers. But I'm not personally so worried about that. It's going to change what people do. The skills you use might be different. The job tasks you have might be different. More people than ever will work with GenAI in their day-to-day work. In fact, the adoption of GenAI in the economy has been faster than just about any recent innovation.

It's going to change how people do their work, but it's not going to put people out of work. It's just going to change what they do. That'd be my prediction.