What “Good” Looks Like: Principles for AI-Enabled Banking Design

“Artificial intelligence” – the buzzword of the last three years since ChatGPT took the world by storm in November 2022. Every process needs an AI solution, every senior meeting has an AI-based topic of discussion, and every solution that doesn’t include some form of predictive model is stuck in the dark ages. What a futuristic time we live in!

However, in reality, many of the AI-based solutions being thrown at problems are inherently neither artificial nor intelligent at this stage. We have chatbots that allow customers to resolve more queries than ever before online. Real-time pattern detection can analyse millions of transactions to spot anomalies and block threats before they happen.

But, in essence, these are all simply examples of big data analysis, a technology that has been around for decades. Granted, these datasets are larger than we have ever had the capability to analyse before, but this “new” technology being employed is not revolutionary or offering anything we haven’t already seen.

So, what needs to change? Well, I am not saying that the processes mentioned previously are wrong! In fact, the value added to the customer is significant. However, banks should start incorporating AI into their basic foundations and building blocks, rather than directly aiming at the customer straight off the bat.

In my experience, systems are often legacy systems and do not speak to one another. This disconnect is addressed manually by human beings. Employing AI at the customer finish line, when the rest of the race is built on outdated processes and systems, is setting the journey up to fail from the start.

Good looks like back-office processes and procedures that work fluidly and flawlessly. It means workers who can utilise artificial intelligence to complete their tasks efficiently and effectively – all ultimately resulting in a better quality of output at the customer “finish line”.