Skip to main content

Banks should be adopting AI... or should they?

(Image credit: Image Credit: Geralt / Pixabay)

AI is the ability of a computer, or computer program, to perform tasks that have traditionally required human intelligence. It’s far more than a chatbot. AI applies machine learning, deep learning and other techniques to solve problems that traditionally only human intelligence has been able to solve. Most banks are a long way off being able to use AI to the full. Before concerning themselves too much with AI for the sake of innovation, banks should first focus on truly automating more of their core systems and processes.

Business Process Automation (BPA) can bring several benefits. More than a decade after the financial crisis, and now dealing with the aftermath of Covid-19, as well as rising competition from challenger banks, incumbent banks are under huge pressure to bring down costs and increase revenue. Better automation and redesign of processes, when done correctly, can really help. 

First, getting BPA right can reduce overheads. The bottom line is that machines are cheaper than people, but the goal here is not to automate people out of jobs. The aim is to redirect resources so that staff can focus on more creative or value-added tasks that will lead to new business for financial institutions, while computers handle time-consuming routine tasks.

Similarly, BPA can help drive new revenue streams. It’s no longer good enough to take months to implement new, digital products and services. Banks need to be able to design and deliver new products and services quickly, and those products and services need to be well-tailored to the individual. 

Furthermore, automation can enable banks to implement real-time, straight-through processes which can address a customer’s expectations and needs. For instance, rather than simply having a chatbot that directs a query to the relevant human agent, automation can help solve a customer’s problem directly.

BPA is more than just automation 

But what is holding banks back from achieving these objectives, or achieving significant success or business benefit from automation? Simply put, they are not approaching BPA in the right way. In most cases, when someone looks at a workflow, the common mistake is to look at what exists, and try and convert manual tasks into digital ones. This doesn’t work because while some manual processes get removed, in the transition to digital often new ones are created.

Most banks are not making the full use of technology to fundamentally reform and improve their processes. While many financial institutions wish to automate their processes, they are making incomplete technology decisions which will lead to incomplete business processes — and the full benefits of BPA will remain out of reach. 

Many banks are making incomplete decisions, overestimating the capabilities of legacy systems and not fully analyzing what needs to be changed in different systems to really enable seamless and automated processes. One example is developing new channels. Quite often a bank will develop new, modern channels only to find out that they need to build an enormous middleware layer with business logic to provide the intended functionality and UX on top of these legacy systems. This leads to duplication of business logic between multiple systems and at the end causes even more problems.

How banks should approach BPA 

What should banks be doing differently? This is not a technological problem; it is a behavioral one. Banks must commit to automation by considering fundamental changes to their processes and workflows. A major stumbling block for most organizations is that they don’t go far enough when automating: they will digitalize a process, but still require a human to press a button to start that process. Automation should be about linking systems together, so that they can trigger one another and drive forward. 

Banks which decide to commit to BPA should follow a three-step process:

First, they must conduct a deep analysis of their existing processes. How many steps are involved? What data is needed? Does this need to be manually input, or can it be automatically drawn from elsewhere? 

Second, check the technological requirements. If you wish to automate a process, check if your current system has the capabilities to do it. Perhaps you need to build an API, or invest in a new system? 

Finally, create a first version, test it, then iterate. The key here is to think digital first, look at how customers behave and think how you can make their lives easier. By doing this, banks will design automated systems that truly address customer pain points, rather than simply recreating manual processes digitally. The version created at this point will not be the finished product; it will be an MVP. This is another common mistake - seeing this as the finish line. Automation is an ongoing process of testing and improving, so be open to changing things along the way

What role AI could play 

As for the role of AI, there are plenty of examples of where it has the potential to aid the resulting automated workflow. For example, checking and converting customer documents, or verifying a customer’s ID using facial recognition. But if banks don’t get the workflow automation right, first by truly removing the need for human input, the benefits of such AI add-ons will be negligible. 

However, intriguingly enough, where AI also has a big role to play is in designing the workflows themselves. Through a technique called process mining, AI can detect the root cause of ineffective performance in a business process, for example uncovering process bottlenecks, or by assessing the profitability of the various decision points. With these AI-driven insights, corporate efficiency can then be enhanced by designing potential AI solutions to intervene right on the dysfunctional link in the process chain.  

Certainly, AI has the power to transform banking automation, and the technology is being trialed in other banking areas, too, such as detecting fraud, preventing cybersecurity attacks, and even providing tailored recommendations to customers. 

To get to that point, banks need to build the foundation first. That means adopting new behaviors, being more open-minded, committing to a fully automated and digital approach, and taking a more critical look at their current processes. What works and what needs to be thrown away? Rather than simply digitally replicating what has come before, banks need processes fit for the modern day.

Ove Kreison, Chief Technology Officer, Modularbank

As Chief Technology Officer, Ove ensures that Modularbank’s products and services lead the market in terms of flexibility and technology and that they match the evolving needs of companies offering, or planning to introduce banking and financial services.