Banks and other financial services institutions face a daunting array of digital challenges. In this interview, Alan Hanson – Partner and Head of Americas Banking Financial Services at ISG, the global technology research and advisory firm – talks about what some firms are currently doing to meet those challenges.

You wrote an interesting post a while back about the impact of driverless cars on the automotive sector. Could you give us a similar 50,000-foot view of artificial intelligence in financial services? Is there a driverless bank in our future?

A driverless bank – that’s an interesting notion. I think some people might say they’ve always been driverless? I will say consumers definitely have a desire for a much more consumer-focused experience with their financial institutions, less people-to-people and more people-to-application. Why should dealing with my bank be harder than dealing with Amazon? 

Will the traditional financial services firms be able to hang on to their brands without that person-to-person connection? Or was that never as important as people thought?

Well, to a certain extent, there is a negative perception of legacy banks. In fact, some have been trying to shed that and cultivate an alternative brand identity, maybe through acquisition – though it’s not clear how effective this is. I was at an event where Dan Schulman, the CEO of PayPal – itself hardly a stodgy legacy institution – was talking about their services, and a relatively young person in the audience stood up at the end and said, “I love this product called Venmo. When are you guys going to become more like them?”  And Dan had to inform that person that, well, PayPal owns Venmo. It is Venmo. 

But yes, I think many banks do have a lot of brand loyalty, which has a lot to do with credibility and trust. And you don’t want to get rid of that. Perhaps more important, banks have a tremendously large customer base. That’s why, in the U.S. at least, so many of the Fintech startups have looked to become partners with banks.

It used to be that consumer preferences almost didn’t enter into banking. Maybe you could entice consumers with a set of steak knives – Or a toaster.  

Right! Let’s not forget the toaster. But now you need to do more. How well are banks doing in attracting customers?  

It’s a major challenge. A colleague of mine came to me the other day and said he had just bought an exchange traded fund through a company called Robin Hood. It’s a Silicon Valley start-up and the notion is you don’t pay any commissions on trades. That sounds enticing, so he made his first trade, liked it, and didn’t have to pay a commission. But I told him, somewhere along the way you’re paying for that, and it turns out Robin Hood makes its money on the float. Now, he’s a sophisticated guy, and he still bought their pitch. If you’re Schwab or Fidelity, you’ve got to be out there positioning yourself against these kinds of offer. 

Are Silicon Valley’s “Merry Men” the only threat?

No, they’re only part of the challenge. A Wall Street brokerage firm with a very large team of wealth management brokers recently started doing some robo-advisory work with Kensho, the machine learning firm, and they tried to straddle the marketing and say, “Hey, if you want high-touch service, we have that, but we also have a robo-advisory service.” It’s a tricky place to be – these guys not only have to appeal to the consumer, but they also have to reassure their human rainmakers that there can be some peaceful coexistence between those two business models.   

What tend to be the biggest internal obstacles that financial services companies face in making this digital transition?

Number one is being able to move fast enough. It’s one thing to put up a mobile app that’s at the outer edge, it’s another to build a digital transformation that reaches to the middle and the back end of the system. Doing that and doing that quickly is difficult. 

The second thing is handling legacy IT. Institutions often have thousands of applications, and those applications cost money to run – companies talk about their “technology debt.” The applications are not all digitally enabled, and most are not in the cloud yet. Trying to spend the money to make that conversion while you’re trying to build brand-new capabilities isn’t easy.

The last obstacle is just having enough enterprise agility – making sure that you’ve got the right technology people coupled with the right product people – so you can bring offerings to market very, very quickly. Each one of these is a not inconsiderable challenge.

Are the banks able to get the right people?  

You see them hiring from Silicon Valley and other places to get the skills they need. They’re also trying to make the work seem more enticing – setting up innovation labs that are focused on products, technology, and new offerings, that are insulated from the parent bank and the chore of trying to rationalize the legacy systems. I think they’re starting to succeed.

Is compliance the big barrier?

The barrier is not so much compliance for banks as coming up with a deeper vision of where they want to go.

We see that often companies go down a road of focusing purely on automation of fixed and predictable processes. That yields some good results initially, but when you start looking at processes that are more complex, like underwriting a mortgage loan, or claims processes, or anti-money laundering measures, you run into a roadblock. We just did a study looking at where 360 companies are with respect to automation, and we found that less than 10% are building automation all the way through their enterprise.  If you began with a view that says, let’s look at this entire set of technologies as one process, then you don’t run into roadblocks once you get running. 

Of course, getting people to change isn’t easy. There’s an old Buddhist saying that in the beginner’s mind there are lots of possibilities but in an expert’s mind, there are very few alternatives. A lot of these companies have people who are experts in various areas. And they are bound to think, “Okay, I need large teams of people to execute a complex transaction,” as opposed to saying, “Hey, I need small numbers of experts who can deal with the exceptions, and then I can do all the rest with machine learning and automation, and this will lead to a better, more effective outcome.” Until you unlock their minds and say, “Hey, maybe there’s a better, simpler, and different way of doing this,” you won’t get anywhere. 

Banks always used to have big in-house shops, but between Fintechs, SaaS, and offshore developers, they have had to become more outwardly focused. Are they getting used to working with external partners?

They are getting much more comfortable with that now. Take the attitude toward the cloud. Two or three years ago, most financial institutions would tell you that they, because of risk or regulations, they could not move to the cloud. That’s changed pretty radically in the last 12 to 24 months. Organizations like FINRA, which is responsible for regulations and compliance, have themselves moved to Amazon Web Services. FINRA now has 60 million client records in the cloud. Now most banks are saying, “Hey, wait, if we can get benefit out of running some of this in the cloud, why not do that?”  

It sounds like the banks are getting more daring than they used to be. Are there any other new technologies where they’re starting to think, “Hey, we can do this”?

Yes, blockchain is a good example. There’s now no shortage of consortia that have developed around blockchain and many, many financial institutions are participating in those consortia, and working on blockchain use cases in their own labs. The same with “backbone” institutions such as DTCC: If you look at their public proclamations about blockchain from, say, three or four years ago and compare them to today, it’s night and day.

Do you have any thoughts about how legacy firms can become more agile?

It doesn’t all need to be agile. Some things have to be containerized and updated every single day. But there’s other work where there’s no need for this. Not everything needs to move at one speed. But it can’t just be bimodal, either. It’s not like there’s some stuff that’s fast and some stuff that’s slow, and that’s that. You have to be able to adapt.

You also need to re-imagine the processes that you have. If you just add technology for technology’s sake, then you won’t be terribly effective. Maybe you’ll develop an Alexa-based app to help somebody do some voice commands, and then turn around and realize that for privacy reasons, people don’t want to bank that way. You’re better off looking at a whole process and asking, “How can I make this better and more effective using the technology that’s available to us?” For example, the mortgage process is complex, and if you understand the mortgage business then you can come in and ask, “OK, how can I use artificial intelligence for underwriting, how can I use blockchains for the title process?”    

We talked earlier about the faceoff between robots and high-end brokers. Do you think people will still have a role going forward?

I think people may be overly concerned about that “driverless bank” paradigm we talked about earlier. They have a notion of massive layoffs ahead, but most institutions aren’t laying people off wholesale simply because these new technologies are introduced. We’ve worked with more than 100 companies on automation and artificial intelligence, many of them financial services companies. What we find is they often wind up shifting the kinds of work people are doing to higher value tasks that make the institution more productive. Financial institutions are always going to need people to make complex decisions. It’s not as simple as turning machine learning loose on a company. Even in designing the automation, you will still need people to guide the process. How are we going to do this? Where are we going to take this?