Somewhat simplified, this picture shows how a neural method and artificial intelligence works. We ask it: “What does this remind you of?”, similar to when people recognise everyday objects when looking at clouds. This neural network is trained on a data set, with a lot of animals (particularly dogs), so those are the motifs the artificial intelligence recognises. Had it been trained on architecture, it would probably “see” pavilions and bridges everywhere. The network then responds something like “I think this part of the image looks like a pair of ears, and maybe a bit of fur there”. Thus, the technique provides an overall glimpse of the main features noticed by the neural network. Image manipulation: Kristian Tølbøl Sørensen, Alexandra Instituttet Photocredit: Johny vino / Unsplash.

The Wave of Automation

The revolution will be ­automated and democritised. Artificial Intelligence is here and it is making financial ­services easier, cheaper, faster and more personalised.

The emergence and increased accessibility of Artificial Intelligence (AI) means big things for fintech. Banking- and finance-related complexities and problems that humans have had trouble addressing are now easily solvable through the use of software and algorithms.

The banking and finance sector is being revolutionized, as AI “toolboxes” are much cheaper to acquire and implement. With the help of AI, banks, NBFCs and fintech startups can create products that are more personally tailored to customers, at a lower price. Through detailed analysis by algorithms and AI, banks can better cater to each customer and cut costs at the same time.
Bill Gates once said, “Banking is necessary, banks are not.” According to Anders Kofod-Petersen, vice-director of the Alexandra Institute and a professor of AI, he was right.

“We have needed banks until recently, because doing clever banking requires personnel. But if we can replace this personnel with algorithms, we don’t actually need banks anymore.”

For example, the process of calculating credit and processing loan approval becomes much faster with the help of AI tools. It’s a simple problem for a machine, but a very difficult one for a human. There’s also a difference between calculating someone’s credit rating and being able to accurately predict whether that person can pay back their mortgage. With the help of AI, these answers can be obtained in less than 12 hours.

“You can’t get to know all your customers, but with the toolbox you can calculate this sort of thing pretty easily and personalize services with greater ease,” says Kofod-Petersen.

Fintechs are adapting rapidly to this changing landscape of more accessible open-source AI by offering customer-friendly products and services that stream­line the banking process. If banks don’t catch up in their utilisation of AI, fintech startups will win big, as they will cut operation costs, meet customer demands and expand market reach.

“Clever algorithms are winning. The discussion about whether you need human insights for this sort of thing is over: you don’t,” says Peterson.

For historical reasons, many financial services have relied on people-solving tasks; however, these tasks can now be handled by software. There’s an ongoing debate in investment: You can pay some banking guy who can outperform the market, or you can put in funds that follow an index. Currently, the probability that an investment banker can outperform the market is about the same as the probability that a monkey with darts can outperform the market.

Today, AI is a toolbox, just like a toolbox you’d buy at any DIY shop. However, this toolbox solves digital problems. Recently, many of these tools have been made accessible because there’s a lot of software that anybody can access. So, the AI toolbox has become a lot more accessible and much cheaper. It boils down to this: Five years ago, you had to build your own toolbox; today, you can find it online.