The job title "head of emerging technology" is freighted with self-assurance; it seems to imply that someone holding such a position should be able to peer, Nostradamus-like, through the veil of futurity and bring back knowledge of the shape of things to come.
Peter Memon, who now holds that title at the global financial consulting firm Synechron, makes no pretense of having prophetic insight, only long experience. And indeed, he has strung together high-impact banking jobs like pearls on a string.
Over the course of a 25-year career in financial services, Memon built front-office trading systems at HSBC, was in charge of architectural development at the ill-fated Lehman Brothers and managed trading analytics data at Barclays. Most recently, he served as the head of a research-and-development team that was designing next-generation technologies at JPMorgan Chase.
American Banker spoke with Memon shortly after his new position was announced to pick his brain about machine learning, data analytics, the future of retail banking, blockchain technology and the "hyperregulated" environment banks have been languishing in since 2008. What follows is our interview, which has been edited for length and clarity.
You've been in finance for a long time. What drew you to this job?
PETER MEMON: The opportunity to do what I do on a bigger scale, across more clients, certainly enticed me. The ability to get out there and speak to people and understand more of their problems and find solutions to those problems is at the very nature of what I love to do. As an engineer, I love to find unique solutions to problems. I don't want to always solve the same problem in the same way as I've done it a thousand times before. And I'm very inquisitive about new things that are out there. Synechron is growing like crazy, and coming here was a totally natural segue.
Financial institutions have traditionally not been good software companies. With the rise of fintech startups, does that have to change?
Financial institutions for a very long time were actually very good software companies. If you go back to the 1990s, financial institutions were highly, highly innovative. They had to be, because most of what they were developing didn't yet exist—everything from their own middleware technologies to their own database technologies. They were doing things like high-performance messaging technology. That was all done in finance long before any of the other [industries] did it, because we had to do it.
I do think it has changed quite dramatically over the years. Post-2008, innovation has certainly fallen by the wayside. Obviously, there are different requirements now, and it's the regulatory requirements that drive a lot of stuff. But we're changing again. And I think the banks ultimately are going to have to change, and I do think they're going to have to become much more software-company-like if they want to compete.
How much should banks worry about fintechs drinking their milkshake?
Think of what fintech companies are good at. They're good at developing a product. What they're not particularly good at, most of them, is becoming enterprise-ready, because they don't have that level of experience. In terms of talent, the talent at a bank will rival [that of] the startups. There are many, many, many highly talented developers sitting in banks. I do think banks ultimately, yes, are going to have to evolve and be able to develop quickly, because the technology is changing. And if they want to compete—especially for the younger generation [of consumers]—then they have to. But that's going to take some time. You're trying to turn an aircraft carrier. For a bank to try to become agile like a small fintech company, that's not necessarily going to happen. But do they need to be more agile? I think so, yes. They need to go back to the way they used to be.
How can a bank achieve that transformation, especially given today's regulatory environment? Is it launching startup accelerators? Is it partnering with fintechs?
That's all part of the equation. Obviously it helps. Again, banks are where they are because of the regulatory constraints. That said, in many ways the regulatory requirements can be used to enable innovation. They really can. And I think a lot of people forget that. But can banks make profound cultural changes on the IT side? Certainly they can; absolutely. And they can do it in a way that, even with the regulations that are in place, allows them to continue to do innovative stuff.
What are the biggest technology challenges facing banks today?
The No. 1 thing I think banks have to do today, and many of them aren't doing particularly well, is to manage their data effectively so that they can monetize it. If you look at the retail side of the bank, they're more sophisticated at that, because they've been doing it for a long time—and retail data, in many ways, is a little more simplistic. But coming from an investment banking background, where the data is much more complicated, I think there's enormous room for growth [when it comes to] managing that stuff better so that you can actually use it, monetize it, on a larger scale. It's a nontrivial problem, given the enormous variety of data that's there. The investment banking side has maybe 50 trading systems, maybe more—one for each asset class or product type—each having its data modeled differently. Being able to effectively monetize that data, that's a much harder problem to solve, and I think they've got to solve it.
[Trying to unify all the systems] is the wrong answer. That was the traditional, top-down approach: "Let me make it all look the same." That's like trying to rebuild a 747 while it's still flying in the air. Ultimately it becomes a massive tax on the organization. I think there are far more efficient ways of doing that. Tools like data virtualization, which allow you to build dynamic models and rapidly evolve those models without having to change the underlying data constantly, allow an organization to be much more agile. Or developing things like views on demand—views that are specific to a problem rather than one global view that's [supposed] to solve all problems. Not all products are the same, or even remotely close to each other; an equity is not the same as a complicated derivative product. You can't model them the same. They look very different.
What emerging technologies are you most excited about in financial services, and which do you anticipate spending a lot of time thinking about and advising financial institutions on?
That's an easy one. Machine learning and AI are going to be the core for banks in general, along with advanced analytics. What you're seeing now is an enormous demand for analytical services. The way it's done today is kind of the craft methodology—it's like [handmade] furniture, where you have your craftsman building a one-off for you, which is a slow, painful, expensive process. Given the demand for these types of services, I can't help but think that the automated generation of models, or even crowdsourced model generation, is going to be the wave of the future.
There are a number of tools out there today that facilitate that. So we've now effectively commoditized model generation and made it available to people who are not PhDs and data scientists. I call it "AI for BI," having artificial intelligence available in tools with business intelligence interfaces. I can make very sophisticated tools available to somebody who is [merely] comfortable using a spreadsheet. That's powerful. I have individuals who want to ask maybe very sophisticated, complicated analytical questions of maybe a relatively small data set, and all of a sudden they can start doing that. That, I think, is going to be a true game-changer. The tools out there now are the first generation; you're going to continue to see them evolve.
Can you give me an example?
Today, if I'm a salesperson at a front-office trading desk, and I want to analyze data for whatever reason—maybe I want to understand the trading patterns of a series of clients—what am I going to do? I'm going to go to my internal data-science team that maybe has 20 to 25 data scientists the bank has recruited, who probably don't really understand my business because they just don't have the expertise, they don't have the experience, they've never done finance before. But I'm going to try to explain my problem to them. A problem that I, as an end user, am intimately familiar with. So I'm going to help them with the data and explain the problem, and they're going to go off and develop some statistical model for me. They're going to test it, they're going to refine it, and then we're going to try to put it into a production environment.
That might take six months; it might take a year—if we're lucky. So they've developed some code that now has to be integrated into some system I use every single day. Very inefficient. Now I have a handcrafted model that, over time, has to be recalibrated, because the exchange data changes, the patterns change; it was tested with one set of data and now it needs to be tested with a different set of data. That requires me going back to that group and saying, "This model needs to be recalibrated."
As opposed to where things are going to go. There are tools now that will look at the data and will automatically, through a series of algorithms, generate a model for me that has the outcome I want. They will sit in the production environment; they are accessible via API calls; they don't require me going to anybody to have them made. I could, using a simple user interface, generate the model, generate the target variables, do all of that stuff on the fly, and get it into a production environment very quickly. Something that might have taken six months to develop, I can now do probably in a day or two.
So I've now made advanced statistical models available to people who previously would not have had access to them. Every place where I have data for which I could potentially come up with some kind of sophisticated statistical model that's going to make my life better, my customer's life better or optimize something, and my turnaround now goes down to a day or two—that's powerful.
Do you think cryptocurrencies pose a challenge to banks, or will they be integrated into banks' businesses in some way? Is it blockchain technology that's truly interesting rather than these digital assets?
Blockchain technology itself will openly be integrated into various functions in banks. Once some of the technical challenges are solved, there are plenty of places where blockchain technology will apply. It's going to take some time, but I absolutely see that. Will cryptocurrencies be successful? That I don't have the answer for. I don't know. I think Goldman Sachs entering the fray [with a bitcoin-trading unit] is fascinating. And it's funny, because Jamie Dimon has a 180-degree [different] view on this.
Banks chase alpha; they want yield. They want to make money. And certainly right now in the cryptocurrency space there is the opportunity to do that, as highly risky as it is. But will it evolve into something that really becomes big? I don't know yet. But the underlying blockchain technology will absolutely, positively take off. There are so many places where it applies—any place where you're going to do reconciliation—bond reconciliation, derivative reconciliation. In all of those places it would be immensely beneficial, especially when you're cutting across numerous firms and having a common book makes sense.
What could regulators do to allow emerging technologies to flourish and benefit banks and their clients and customers?
Regulators are typically behind on a lot of these issues. It's not their wheelhouse. The technology is changing far quicker than they can absorb. There has to be the ability to absorb these changes into banks a lot faster than we are doing today if they're to have a big impact. And we are not there. The regulators typically lag.
Do you think regulatory concerns are the main reason banks have been slow to adopt new technologies? Or is it simply a culture issue at banks or something else?
A lot of the bank culture issues are driven by the regulatory stuff. Since 2008, the banks have largely been living in a hyperregulated environment, and that's driving a lot of the decisions that are being made internally. After 2008, the cultures changed because the regulatory environment changed so dramatically. It's making the banks wary of change, so by definition they're more risk-averse.
But it's changing. I see banks coming back to wanting to embrace change, wanting to embrace new technology. They see the need to do this, and it's just kind of happening on its own. The banks that I've worked at are now embracing some of these things; many of those at the top do believe it has to happen, and want it to happen, and are facilitating it internally. I think it's going to be a long ride, but it's happening. The market is a beautiful thing. And the market will drive you to make decisions. When there is a lot of external pressure, and a lot of potential competition, you're going to change.
What is a potential challenge about which banks aren't thinking enough?
There's nothing stopping an Amazon or a Google from becoming a bank. And they're sitting on massive piles of cash. They have cultures that are deeply technical and comfortable with analytics and big data. That's your competition. But if I was an Amazon or a Google, why would I want to walk into a regulatory nightmare? It isn't that they can't get a [bank] charter; if I'm sitting on $250 billion of cash, I'll get a charter. But why would I want to walk into a hyperregulated environment? That would be my first concern if I was them. But the possibility is there; nothing is stopping them from doing that. They are very well-suited to do it. That's what would keep me awake at night if I was a bank: "Which of the big guys want to be in my business?" Not the little fintechs.