3 key considerations for responsibly adopting AI in banking

In case you’re somehow yet unaware, the development of artificial intelligence (AI) continues to take the world by storm, and that certainly applies to the financial sector. We’re now situated firmly in the “hype” phase, where we’ve seen the seemingly limitless potential applications that other banking leaders are already beginning to leverage. Now all that’s left is to see whether regulation and red tape will hamstring banks before they even get properly started with AI adoption.

But never forget, AI is far from a new topic. In fact, the term was coined by a group of researchers at Dartmouth back in 1956. And while the gestation of this world-changing tech has taken quite some time, we’re finally at a crossroads marked by the explosion of the AI ecosystem, which is now focused on accessibility and cost efficiency — making it easier than ever to consume and deploy. Rather than locking the tech behind proprietary barriers, the tech leaders of the world have united to accelerate AI’s evolution, meaning the banking industry will feel its full impact sooner, rather than later.

In the following blogs, we’ll dive deeper into the business and customer applications of this revolutionary technology, but for now, let’s set the scene — giving you some things to keep in mind and hopefully allaying your concerns about AI.

Let’s be honest — AI isn’t new, but the potential banking applications are. That means there’s a huge education curve for all of us, and that includes regulators. So if you’re worried about the adoption of AI in banking, you’re not alone. Many banks just like yours are waiting until regulation is more firmly in place to decide whether this kind of tech is worth investing in and to what extent.

As we speak to our customers, we’ve noticed a few commonalities in the perception of AI and its applications. For one thing, most bankers are very comfortable to start applying AI in service of efficiency, using it to enable automation and empower scaling, all while eliminating waste and human error. Behind the walls of the bank, so to speak, seems to be a safe zone for many institutions, and that’s becoming obvious from the early adopters who are already rushing new internal products and services to market. In fact, McKinsey estimates that gen AI could add between $200-340 billion in annual value to the global banking sector, or 2.8-4.7% of total industry revenues, thanks largely to increased productivity, as demonstrated in the chart below.

However, there’s a level of hesitation on the customer-facing side of things. With the regulatory environment yet undefined, it’s potentially risky to bet the farm on tech that might not be marketable — or even usable — in the years to come, and many banks have adopted the “wait and see” approach here. No one wants to be the first one to invest millions on a proposition that doesn’t pan out. Just look at the embarrassing launch of Google’s AI overview, which has proven to give false, misleading, and dangerous answers to common queries.

But let’s make one thing clear — you can have your doubts about AI, but this wave of tech is coming, no matter what you do. If you can set aside your reluctance, the question then becomes how your bank can responsibly and cautiously begin implementing — or at least considering — AI so you’re not left high and dry when the ecosystem is more mature. Let’s take a look at 3 things you should keep in mind while beginning your bank’s AI journey.

First of all, the worst thing we could do with AI is add it to yet another series of silos. Even though the tech is highly advanced, it all means nothing if it’s used in service of yet another point solution. So don’t merely jump on the AI bandwagon and rush out a new series of “innovative” products and services that exist in a vacuum. Instead, try to find a way to use AI to drive value, both for your bank and its customers, fundamentally reimagining how you can leverage this tech to provide superior experiences across your value chain.

Naturally, adopting AI entails some risks, but consider the fact that you don’t have to use it to completely automate the customer’s entire financial life cycle from Day One, nor should you, due to regulatory concerns. Rather, you can harness it for generating personalized, data-driven recommendations which customers can then choose to take action on. For example, AI can analyze a customer’s financial profile, risk appetite, and stock options, then provide critical insights into their spending habits, create a pre-approved credit card offer, or suggest paths for improved financial wellness. And once you’re more comfortable — and regulation is figured out — you’ll be able to do increasingly more with AI as the years go by, but it all starts with creating a solid foundation.

Lastly, don’t forget, AI — and all tech, really — should always be used in collaboration with your employees, enhancing their abilities rather than fully replacing them. Keep in mind that digitizing your old ways of working isn’t what customers are asking for.

So as you’re brainstorming ways to implement AI at your bank, use this tech to augment your employees, rather than supplanting them. By complementing and extending their abilities and decision-making, you’ll be able to do things like speeding up credit risk assessment and analyzing data to provide more personalized financial services.

And speaking of these potential uses — in the next blog, we’ll explore the business applications of adopting AI at your bank, including everything from customer acquisition and retention to boosting share of wallet. After that, we can move on to the customer applications so you can have a comprehensive view of all that you stand to gain by responsibly adopting AI at your bank.

For more information, check out our Banking Reinvented podcast, where Backbase Founder/CEO Jouk Pleiter dissects similar topics alongside Tim Rutten, EVP/Chief of Staff, and other digital leaders. Stay tuned as they chat about everything from progressive modernization to decomposing your bank’s complexity.