Financial technology (FinTech)— one of the fastest-growing sectors in finance— is transforming conventional business institutions. But what does this mean for big banks?
According to Sanjiv Das, Ph.D., the keynote speaker of the inaugural Gabelli School of Business Fintech Conference, by 2020 at least five percent of all economic transactions will be handled by artificial intelligence (AI).
“The banks are saying, ‘what can we do with modern technology to actually monetize the data that we have?’” said Das at the March 16 conference that focused on blockchain, cryptocurrency, machine learning, textual analysis, risk management, and regulation. “The prognosis is that any bank that doesn’t become a technology company is probably at risk.”
A Santa Clara University professor of finance and business analytics, Das identified 10 areas where FinTech is gaining clout, including fraud detection, cybersecurity, deep learning, and personal and consumer finance. He cited mathematical innovations, hardware, and big data as game changers.
“The fact that they now have mathematics that actually allows us to include very large-scale models with millions of parameters is absolutely key,” he said. “You have to feed the beast, and the beast eats data.”
As the FinTech space expands, Das said, financial institutions are faced with an important question when it comes to talent pipelining: Should they train their in-house engineers in finance or teach their finance professionals technology?
“My bet is that you can take finance [professionals]and teach them technology,” he said. “Everything has become so commoditized that it’s actually very easy to do this with the tools that we currently have.”
Still, to excel in the sector, professionals need training across disciplines, he said. “You’re going to have to learn something about behavioral psychology, cognitive science, computer science, and statistics.”
“If you want to get under the hood with of all of this, the two skills you’ll need to learn are linear algebra and statistics.”
Enhancing customer-service
Some banks are tapping into conversational AI and chatbots that assist customers in managing their personal finances.
This month, Bank of America launched its virtual financial assistant Erica, a chatbot that helps users with bank-related issues such as making payments, checking balances, and reaching a savings goal.
Das said that chatbots like Erica aim to enhance customer-service experiences in financial services.
“When you call customer service, there is a huge variety in the quality,” he said. “You might get somebody who knows what he’s talking about or you might get someone who is one week on the job. If you can replace those people with a chatbot…you’re going to have much better service at a very low cost.”
AI has proven good at predicting things where data are stationery–for example, detecting cancer through cells that don’t change. But AI is less effective at making successful market predictions, Das said.
“Market predictions is a tough problem because markets are not stationary, so we need to figure out better models,” he said.
While experts have argued that machines will never outsmart human intelligence—even though they learn from experience— Das doesn’t think humans beat machines in every domain.
“Humans learn from experience too, but we can’t do a million games over a weekend. The machine can. It’s faster learning and it’s more accurate.”
“What humans are better at is explaining why they made the decision whether it’s wrong or right,” he said.
