To keep up with the pace of business and the security threats in today’s environment, financial institutions need to improve efficiency and develop innovative safeguards to manage risk. The advance of artificial intelligence (AI) and digital assets makes that possible, improving performance while reducing processing time and costs. Though some of these advances are already in use, the level of sophistication will progress to such an extent that the banking industry will be set up very differently in the next decade.

In an interview, Henry James, founder and deputy CEO of Fincross International, described what he called a “helicopter view of banks using AI at the early stages.” He explained that among major banks there is already an understanding that AI will be incorporated in multiple areas, ranging from how to better manage the risk of financial markets, data security and compliance issues.

“You name, it,” he said, AI can be applied to “pretty much any risk that a bank faces.” He also believes that there will be major growth in such uses.

The AI Future will Take Time to Arrive

Currently there are some limits to AI adoption due to the the fact that experts in the field are still “very scarce and very expensive.” Consequently, deployment requires a large investment. That cost coupled with uncertainty and a reluctance to let go of legacy systems is what makes some banks still hesitate to fully embrace AI at this point.

He explained, “You can never be sure what the outcome and the success and accuracy of that AI will be” down the road. Though the nature of machine learning is to advance over time, “it requires continuous refinements” that will take into account “new threats and risks that the AI has not previously encountered.” Also, bringing in sophisticated AI solutions means abandoning the “old school” software setup that has dominated the industry.

We will see that give way, though, as the the future belongs to those who are able to use flexible tech. “I think going forward, the technology stacks are going to have to be flexible, integrating with several other solutions,” James observed. “The level of customization and integration will have to be significantly more flexible than the solutions used today.”

But, one should not expect to see this happening overnight. “The transitionary period will take years to move away from legacy systems to a wider use of fintech and artificial intelligence.” He estimates that arriving at “the next gen of banking” as “the norm will take upwards of 10 years.” (To learn more on fintech, check out What the $#@! Is Fintech?!)

How AI Is Shaping the Present and Near Future

Though James sees the real future of AI arriving further down the road, concerns about cybersecurity are driving banks to find better solutions than the two-factor authentication that has been proven to not be altogether secure. While they are “still better than alphanumeric passwords,” he said, “hackers have found ways to bypass them.”

At Fincross, there is a team actively working on AI innovations in its focus on trading tools for the cryptocurrency market, as it concentrates on digital assets. Those include developing biometric AR for leading-edge tech in reducing fraud. One of its solutions is just about to launch.

For trades involving millions, the bank has pioneered a way to use its app to ascertain that it is indeed the account holder putting in the order. James explained that it works by having the person take a short video in a room of the user’s choice, whether that is at home or in an office, that shows the surroundings and is sent on to the bank. Then when a trade order or withdrawal of significant amount comes through, the bank would request a return to the same environment that they can both geolocate via the phone and match to the identification video via the app.

This extra step would not be required for your average type of fund movement, but, James said, people are willing “to go to additional lengths to protect themselves when it comes to transactions of millions or billions of dollars.“

Adoption of Robotic Process Automation

Banks are also already making use of robotic process automation (RPA). Among them is BNY Mellon, which began deploying bots as a way of capitalizing on AI capabilities to reduce costs and improve operational efficiency about three years ago. Likely others will follow, as RPA is an area of spending that Forrester predicts will grow to $2.9 billion within two years.

The cost savings are significant. In 2017 Reuters reported that the bank estimated an annual savings of $300,000 as a result of the shift from manual processes done by people to the automation enabled by bots. As for efficiency, the bank reported the following numbers:

  • 100% accuracy in account-closure validations across five systems

  • 88% improvement in processing time

  • 66% improvement in trade entry turnaround time

  • ¼-second robotic reconciliation of a failed trade vs. 5-10 minutes by a human

The technology adopted by the bank for this is Blue Prism. In the video below, Dave Moss, CTO and co-founder of Blue Prism explains how the Blue Prism Robotic Automation Software Platform works:

The idea is that the robotic automation powered by AI can bridge the gaps that currently exist in implementing technology that had required human intervention.

Reducing Human Labor

What follows naturally from eliminating the need for that human involvement is a loss of the need for human labor and fewer jobs for humans because the work will be done better and more reliably without them. That’s the spin Sandeep Gawade, operations manager for BNY Mellon in Pune, India, gave here:

Robots are reliable and deliver what they are designed to perform. They’re not affected at all by factors like workload, absenteeism, attrition, stress, or holidays. In fact they reduce risk and improve quality in a controlled environment.

He did also add some of the usual kind of rationalization that more rote work done by robots means more interesting work for humans: “Automation releases our people to focus on decision-making activities. It also eliminates tedium — we hire people with advanced skills to analyze data, and it’s disheartening for them to spend 30% to 40% of their work time on rote tasks. With help of robotics, we can increase their efficiency and help them focus on more productive work, including direct interactions with clients.”

But you don’t need advanced analytics skills to be able to foresee that cutting back 30% to 40% of work time will, inevitably, mean eliminating 30% to 40% of headcount. That’s a major concern that James has for the AI-powered future. “Today’s back offices could be in the thousands of employees,” he observed. “A very large amount will be replaced by AI.” (Another big advancement in fintech is mobile banking. Find out more in The Impact of Mobile Banking.)

Planning for the Future

The fact that there will be fewer jobs available in banks, as well as in other industries that will rely more on AI and less on human labor, is one major issue that requires planning for a sustainable economy. Another is the regulation of AI itself.

“Financial regulators are now having to come to grips with innovations in fintech,” James pointed out, and AI is the area that “will have the biggest impact on the future of any bank.” He expects shaping the regulations for AI is “going to be a bit of a minefield.”

But regulation is necessary because it is possible for AI to be used not just to counter fraud but to perpetuate it. He explained it can be set in a way that “hides and manipulates markets, which is a serious threat to the compliant operations of banks, as well as every vertical that adopts the use of AI.”

It’s a serious problem, James stresses, because when “AI is used to its full potential, it is several times more sophisticated than humans are.” And that is the dual-edged sword inherent in the advance of AI: It is a very powerful force that can increase efficiency, but it also can have negative consequences when directed for nefarious purposes or when it eliminates more jobs than it supports.