PrecisionLender CEO, Carl Ryden, recently sat down with Jim Young, Director of Communications, on the Purposeful Banker podcast to talk about artificial intelligence and machine learning, topics that are front and center at PrecisionLender.
You can listen to part one of their conversation here, but we’ll also be pulling out excerpts from the transcript and modifying them into blog posts. Today’s topic:
The Current State of Bank Data
(Note: While there are definitional differences between artificial intelligence and machine learning, for the purposes of simplicity, we’ll only use the term “AI” in this excerpt.)
Jim: When we’re talking about data sets and the importance of them for AI, how does the banking industry’s data compare to some of the other industries that are diving headfirst into AI?
Carl: A lot of banks’ data was not built for the purpose of distilling it into intelligence. Most is data that’s meant to process transactions. It was built for tactical means: “I’ve got to debit this account and credit this account and make sure no pennies get lost.” That is the core of what they do, but it’s also the most commoditized thing they do. That’s table stakes now. Most bank systems were built merely to handle table stakes.
Over time those systems have been modified to where they can satisfy regulators and pull reports and other things, but they weren’t built to deliver intelligence back into the system. That may seem like a criticism, but there are also valid reasons why the systems were built this way.
Most of these systems were designed and built 20 years ago, when storage was $500,000 a gigabyte; every bit was precious. But now, with companies giving out 15 gigabytes for free, there’s an ability to store everything that might be useful. And there’s plenty of cheap processing power that can sift through and separate the wheat from the chaff later. So there’s an entirely different design principle at work now than was previously the case.
What Happened and How Did We Get There?
But again, the current system at most banks is still focused around transactional data. Banks typically only care about the outcome: “What was the rate? What is the payment that is due?”
How you got to that outcome, and the decisions that were made along the way by humans, is really important data. With AI, we can train a machine to understand those decisions and be able to offer supplemental information and suggestions based off them. With AI you can identify: “Okay, I see Banker A over here actually performs much better.” And then you can get to that causal mechanism so you can find the answer to, “What does he do differently?”
That’s something we integrate deeply into PrecisionLender. We want to understand not just where you ended up, but how you got there, what choices were made along the way, and why. Then, when we see a similar circumstance, we can offer a suggestion about what to do.
Follow the Trail Blazed by E-Commerce
A lot of banks we talk to say, “Our situation’s unique, because of regulators.” There is some uniqueness to the banking industry, but we think it’s overplayed a bit. We see immense value in simply taking things that have already been done in e-commerce, with Amazon and Google and other folks, and applying them into banking.
For example, when you’re using our tool and you’re pricing a loan or pricing an opportunity in the context of an existing relationship, it’s no different than when you’re going through Amazon and searching for big screen televisions and whatever. Amazon knows the transaction you’re in, they know which things you’ve looked at, which things you’ve considered, and which things you might consider next. They also know the relationship they have with you. What are the things you’ve bought over time? What are the things that you’ve rated? What are the things that you’ve said were really good and said were not so good?
When you’re talking about pricing an opportunity to a commercial customer in the context of that relationship, it’s not just the current state of what they are, it’s the entire history you have of how you got to this point that can be brought to bear. That’s something that a computer can do a lot easier than a human, because relationship managers leave and move on to other banks, and new relationship managers come in, and so on.
A lot of the problems in banking are much easier for AI to solve because they are, by nature, data. It’s numbers, it’s balances, it’s finance. It’s not nearly as general as the consumer sphere. Building an all-purpose AI that can answer any question you throw at it – that’s ridiculously hard. When you narrow the fields and you’re only answering questions about commercial banking, the level of difficulty drops considerably.
AI = Amplified Intelligence
For your bank to build a brand and maintain strong relationships with your customers, you will have to have AI. If you don’t, you will be at a severe disadvantage. It’s not about replacing the humanity, but rather amplifying it, giving your RMs a greater, richer context. It’s using AI to see across RMs, across the bank, across time, and look at huge sums of data; not just at what happened, but how it got there, and then putting that in the hands of the RM. It’s acting as a digital guide for RMs, letting them know “Hey, you might ought to know this,” or “Hey, you might ought to consider this.”
Those things will be incredibly powerful, but very soon they will also be table stakes.