“You have to understand the role that product plays in what [the customer] is trying to do — and it better be important — or it’s just not worth the kind of mind share and prioritization of that customer to onboard you.”
Audrey Kim, a venture partner at Oak HC/FT, focuses on growth equity and early-stage fintech investments. The firm has nearly 100 companies in its portfolio, including Cart.com, Kasisto, Noom and Rapyd.
Prior to joining Oak HC/FT in 2023, Kim spent over a decade in business and product leader roles at fintech giants Stripe, where she was general manager, Square Capital, as head of product, LearnVest, where she was product lead.
Becoming an investor was the next logical step in an already storied fintech career.
“I had the great good fortune of running across Annie Lamont and the Oak team, and it really resonated with me to have a team covering global fintech, and to apply a lot of sector knowledge, experience, and beliefs about the future to try to support really amazing entrepreneurs across the board,” Kim told Future Nexus.
Being entrenched in these iconic fintech companies helped Kim learn how to build a product, predict that product’s long-term success, and the depth of impact it can have on a customer’s performance.
Future Nexus asked Kim about how entrepreneurs can best think about customers and how their products fit and AI’s influence in fintech.
This article was lightly edited for length and clarity.
Having worked at both Stripe and Square, how do you impart those “lessons learned” to your portfolio companies?
I have a test that I advise all my portfolio companies to use, which is, can you write the performance review of your buyer? They have three bullet points that justify why they should get a bonus, and your product should support one or more of those bullet points so they can say, “I’m such a hero for bringing this product into our company.”
If you are part of one of those three bullet points, and you’re at the top of the list, awesome, you have made your customer look like an actual hero. If you are not on that top three list, it’s going to be really hard to get their attention, even if your product is, in fact, better than all the current alternatives. They just have limited time and mind share to champion you.
Do you think that we’re still seeing a volatile economic environment?
We view the macroeconomic conditions as constraints, both for ourselves and also for our founders, as opposed to there’s a whole new opportunity set to go chase, because it happens to look like this right now.
We definitely think about that when we think about how to deploy capital and how we advise our founders to think about execution and potential capital markets. No matter what the environment is, you just want someone who’s laser-focused on delivering compounding value to the customer. And if you just do that, you will be able to pair that with sound financial judgment that is constantly informed that the market can change, and that applies to both economic environments and also the mood of the venture industry.
You can go from a time where you know the goalpost is on one end zone, and then it doesn’t just move 10 yards, it goes to the other side of the field, and you have to build both a business model and a team that is resilient to that, because it will happen, and it particularly affects financial services businesses.
How do you view fintech when it comes to the kinds of investments you want to make?
The first generation of fintech, Fintech 1.0 was all about delivering digital versions of traditional financial services, especially on mobile. I’m really glad that happened, but a lot of the juice has been squeezed from that. When we look for Fintech companies of the future, we define it a lot more broadly than, say, neobanking. We like to think of it as investing in software and infrastructure that touches money, and that can look like a lot of different things.
Two buckets that are useful for us are enterprise software that helps manage financial flows for any kind of business, so that looks like supply chain management, ERP (enterprise resource planning) systems, and risk management across the board.
The second bucket that we look at is products that support financial services or money movement. There are payments, and within that, we actually do a lot of stablecoin infrastructure. We were one of the early investors in Bridge (acquired by Stripe).
We also do tax automation with Blue J, for example, using AI. So I think those companies have much broader reach than what would have been called fintech in previous years.
Speaking of stablecoins, how has the recent regulatory landscape affected stablecoins?
It’s just clarity. Great businesses appreciate the importance of regulatory compliance if you are going to tap into the really big markets.
How can AI play a role in the growth of a product?
We do think AI has a lot of impact, and many companies have been funded, but we’re in super-early innings. The thing that we care about most is that the product doesn’t just make things cheaper. It’s not about “can I take the same amount of work and do it with fewer people?” It is about, “Can I take the same number of people and do fundamentally different kinds of work?”
That distinction is really powerful, and ultimately, does that translate into revenue? A lot of the teams that our companies serve are back-office or historical cost centers. There are ways that you can actually turn into revenue generation, and, obviously, that’s really valuable to the customer.
One concrete example is often seen as a kind of cost-cutting measure. You have a bunch of fraud, and if you reduce the fraud, then there’s less of a hit on the bottom line. But for both financial services businesses and digital businesses that have really important things like identity components to them, if you have a better understanding of risk and a much finer-grain understanding of who your customer is, you can actually really lean in to giving them more services. For example, give them more credit, or you can pay them out faster.
Another example was at Square where people with good businesses couldn’t get payment accounts with other payment providers because their credit score was bad. Instead, Square onboarded people, watched their behavior, and used machine learning models to predict if this is a good or bad actor, and then they only threw out the bad actors.
There’s a lot that you can do with AI there as well, primarily because the way that companies implement risks often are to do very simple checks, and then for anything that looks complicated, push it out to a call center or a series of humans. However, you can have large language models both ingest way more data per case and apply it more consistently and increasingly, take feedback really fast, and figure out patterns. You can then turn all that into revenue.
Which areas of fintech aren’t as popular these days?
It really does lead back to what we were talking about earlier about the macro environment. For a long time, the Fed printed easy money, so it was very easy to grow based on a series of assumptions about what was going to happen with interest rates.
Balance sheet-intensive businesses, where the business model is predicated on rates being at a certain level for a long time, are very vulnerable to shifts in macro that you don’t ultimately control.
That said, there is still a ton of opportunity for underserved banking, especially in emerging markets like Latin America, where you need foundations for stable funding. You just need to anticipate and deeply understand that there’s going to be volatility that you can’t control. And you need to build your business in a way that’s durable to that.
How do you see stablecoins as the future of fintech?
It just has to be really compelling in the compounding value it can offer customers. For example, we look at a lot of very early-stage stablecoin companies; maybe they use large language models as a tool, but we think that is actually a pretty profound shift, and honestly, it’s something that we’ve been following for a long time.
If you look at our portfolio, Bridge was even one of our kind of later stablecoin investments, and they were a big part of getting to a tipping point in terms of mainstream buy-in to even the concept of stablecoin infrastructure. When there is a very clear regulatory framework emerging, it’s important to get people on the same page and create a market. Then comes interest from the major institutions, who previously thought that it was a scam.
Previously, when any financial services business wanted to move into a new geography, it took a long time because you essentially needed to rebuild a whole stack in order to fit local regulatory requirements and technical requirements. Fintech businesses are like any business — if it takes a long time to build, you’re only going to go to the biggest markets. Otherwise, it’s a negative.
What’s really powerful and cool about stablecoins is they’re default global, and what that means is you can have a lot of markets, or in the long tail weren’t worth it before, and are now worth it, and there’s a lot of access.
A good example is Stripe just worked with Bridge and Lead Bank to issue stablecoin-linked Visa cards. Now the entrepreneur who’s sitting in Latin America, who used to pay extra on AWS services, can run the business with access to the global financial grid that American businesses have always enjoyed.
