Next month, the BreakThrough Treasury and Finance forum will focus on some of the biggest innovations happening that are impacting financial professionals. And unless you’ve been living under a rock for the past five years, you’ll know that when it comes to innovation in the treasury and finance space, there’s no hotter topic than fintech.
AFP spoke with Laurens Tijdhof, partner at Zanders, who is leading a discussion on how blockchain, artificial intelligence (AI), robotic process automation (RPA) and other advances in fintech have the power to revolutionize and disrupt treasury teams.
AFP: Your session explores how fintech can impact and disrupt treasury. Can you provide some examples of how this is already happening, or how you could see it happening?
Laurens Tijdhof: There are a lot of buzzwords currently that we hear in the market—RPA, big data, machine learning, AI, distributed ledger. But I think these are more than just buzzwords. One or two years ago, everyone was in a phase of understanding what all these new technologies are. Currently, we get daily requests from our clients, asking how they can prepare for the future and how this will impact them. But they’re also opportunity-focused; they want to know what they can do with these new technologies. Some of them are working on proofs of concept and pilots to apply the technology, and some are already applying it.
With fintech, the key is to look at which technology will most likely be adopted immediately, like RPA, or robotic process automation. Maybe you could even call that a new term for something that was already there 10 to 15 years ago—a macro in Excel. That’s also a kind of RPA, but RPA solutions are more advanced; you can connect systems and workflows and automate manual steps. And you can apply it in corporate treasury when the process is already standardized and rules-based. Because if that’s the case, then it’s relatively easy to have manual, operational steps in the treasury process to be automated by robotics. But if the process is not standardized, or if it’s a broken process, then you first need to focus on standardizing your process before you can think about automation with RPA.
AFP: For those treasury functions whose process is already standardized, what specifically are they using RPA for?
Tijdhof: We’ve seen RPA solutions used in the treasury environment for cash application, treasury desktop automation, exposure determination, FX trading, some back office activity—routine tasks that are repetitive, time-consuming, error-prone processes, which demand high accuracy and speed and can be done better by robotic solutions. If you think a little bit bigger about concepts like machine learning, we also have seen companies working in these areas as well. It’s a bit newer, but there you are more dependent on the data. If you don’t have the data, it’s not so easy to apply machine learning techniques. So a lot of companies are currently working on getting the right data in and then having the solutions to analyze it and get predictive information out of it. It’s all about predicting something from historical information. I also see companies using machine learning for cash flow forecasting, and to detect and prevent fraud in their payment processes.
AFP: Going back to your earlier point on machine learning—the most important part is gathering all of the data that you need. What are you seeing companies do to amass that data? What tools are they using for that?
Tijdhof: On one hand, we see consolidation of ERP systems. SAP is a very strong player in that area; I would say about 50 to 60 percent of large multinationals are clients of SAP. Most of them have a roadmap in place to get to fewer versions of SAP, perhaps even one version for the entire company. If that’s the case, then you have a great starting point where all data is in one system, and you could use it for prediction.
The other alternative is connecting all different types of systems to build a data warehouse. Then you have all the data in one central place and you can start doing something with it like machine learning.
AFP: Investors are dumping billions into blockchain/distributed ledger technology (DLT) startups even though very little has actually been produced, particularly when it comes to solutions that could directly benefit treasury and finance. Treasurers tend to be focused on real results; they tend to only allocate funds where they are needed. How do you see blockchain really becoming something that treasury will invest in?
Tijdhof: I’ve seen three use cases that are very promising, but they each have their own challenges. You might be aware of the proof of concept that Microsoft built with Bank of America to simplify the trade finance process. Using a blockchain, they were able to reduce the 15 steps in the letter of credit process to five steps. The error rate is then zero percent, and you cut the time from five days to about five minutes. That’s in a simplified environment, but it looks very promising. If you would be able to get the whole industry behind this conceptual thinking, then the savings would be amazing.
The second case is in correspondent banking. SWIFT has a challenger in Ripple, which created its own network on a blockchain to settle cross-border payments—much faster and cheaper. But the challenge there is that they will never fully replace SWIFT, because it’s owned by the banks. Why would banks give up their solution, which is still very successful? And why wouldn’t they develop their own new DLT-based solution instead of ceding the market to Ripple?
The third use case I see for DLT is in post-trade settlement. You might have heard of Continuous Linked Settlement (CLS), which is also owned by the banks. It was founded after there was a big failure in Germany with intraday settlements of derivative positions. Basically, if you have two financial institutions with major derivative positions, and one is paying out their amount and the receiving amount is not in yet, you might have a big exposure there. So that is why a system like CLS was introduced. But the technology behind it is relatively old and you could do it much more efficient on a distributed ledger. But you would be dependent on the number of financial institutions or corporates participating in an alternative network.
So, long answer to a short question: the technology is superior, but hurdles like adoption by the industry and getting the regulation aligned need to happen before these things replace traditional solutions.
Don’t miss Laurens Tijdhof’s session, Fintech and Treasury: A Match Made in Heaven, at BreakThrough Treasury & Finance, May 6-8 in Nashville, sponsored by Bank of America Merrill Lynch. Register by April 6 to save.