Everything you want to know about how AI is transforming banking but were too afraid to ask

By Alara Basul | 27 March 2017

The rise of artificial intelligence (AI) is set to change the way banks and financial services operate, as well as the way consumers approach their personal banking. Powerful AI can replace humans with machines, improve customer experience and provide simplified cost-effective solutions for businesses.

bobsguide recently spoke about AI’s impact on the fintech sector, today we’re digging deeper to find out how AI could transform banking by speaking to Tom Blomfield, Brett King and Spiros Margaris to discuss their views on AI and machine learning.

What is AI?

AI is the science of creating intelligent machines and computer programs that can replace human tasks. Computer programs have vast capabilities to execute tasks quicker than humans can, with embedded algorithms that leave less chance to make human error.

Robert Smith, Chairman and CEO of Vista Equity partners said at The World Economic Forum: “Since the invention of computers, we have envisioned that computer systems will take the best of what we think and deliver real time solutions that are more efficient. The desire to leverage really hasn’t changed since we were first sparked with that vision. What has changed is that we have scientists, technologies and leaders who have developed new computing platforms such as AI that make this type of computing a reality.”

Tom Blomfield, CEO of Monzo says: “Broadly speaking, the way Monzo is defining and using AI effectively is through supervised learning. We have a big data set which we analyse to determine outliers. A good example of this is detecting fraud. We take a look at the vast data from historic transactions and can identify that a small portion of those is fraudulent. We feed that data into a machine learning model to build a set of rules or an algorithm that will, in future, detect which transactions are fraudulent.

“Our machine learning engine will then very quickly identify fraud and fraudulent transactions and suspend these accounts.”

Spiros Margaris, VC and fintech expert says: “I think almost everyone has some understanding of artificial intelligence (AI), even if the person does not work in the technology industry. I would describe AI as computer intelligence, or to put it another way, machines that try to simulate (or one day to exceed) human intelligence and perform tasks cheaper, faster and better than humans.”

How long have people been talking about AI in banking?

“The ideas around AI and AI as a concept have been around for at least 30 years; the difference today is that we have a lot more “labelled data” available because of the internet. The processing power in terms of technology and computer programs now have the power to run these algorithms. The last two three years is where we’ve really seen a step forward with AI”, comments Tom.

“AI went through natural waves of enthusiasm in the banking industry due to difficulties in finding good and affordable use cases. We now see a very strong phase of AI due to cheaper and almost abundant computing power; and due to the enormous data flood by smartphones, Internet of Things (IoT) and social platform engagements.” Spiros adds.

“However, I am certain there will be no more banking, insurtech or fintech without industry participants incorporating AI as part of their solutions and competitive advantage.”

What are the major areas that AI will make a difference to banking?

Blomfield says: “There are three areas that AI is set to make an impact.

“The first is fraud and fraud detection. The second is credit underwriting; the process of deciding who is going to be creditworthy and why. The way credit underwriting is done at the moment is, very broadly speaking, individuals’ names and addresses are checked with a credit reference agency to acquire a credit score. If their score is over a certain threshold, they can borrow credit. It’s very naïve. There are a variety of indicators you can feed in; maybe your credit appointment should depend on your employment, where you live, who your friends are, or whether you spend your money on Uber or black cabs. There’s a huge amount of data that’s not being used.

“And the third area is personalisation; recommending products and services based on your data. That being said, for these three areas you can actually get 75% of the benefit by simply having the data in one place and running regressions over that data, using traditional linear regressions and writing rules. The biggest problem is getting the data in one place.”

“We are still in the very early stage of the potential of AI in the banking industry. Where I see AI having its biggest potential and impact, besides the automation of back-office processes, wealth management and cybersecurity, is in providing customers with affordable personalized financial services”, Spiros Margaris adds.

“I believe the ultimate banking experience we should aim for will only happen when we achieve a level of AI integration in banking or fintech that feels, acts and is very similar to what Marvel’s Iron Man ‘Jarvis’ AI-version represents. The future AI-assistant will not help build weapons like in the Iron Man movie but will be a perfect assistant to access information and provide recommendations for everything we have data on—and obviously for all of us it will go beyond banking.”

Are people replacing humans with machines because it’s cheaper, or because it fundamentally improves the quality of the product?

“It genuinely improves the quality. The complexity of a machine learning algorithm is often that the output is not comprehensible to a human being. They cannot simply provide three rules that helps you identity fraud. For example, as a human, you can understand a three dimension matrix with three different inputs that might impact the different outcomes. With machine learning, it’s as if you’re thinking in a 90 dimensional space, it’s just not comprehensible. The machine learning algorithms can do things that the human brain cannot compute.

“In our data we we’re seeing about £40,000 a week in fraud, and our human engines brought that figure down to about £8,000 a week. Our AI engine brought that down to between 0-£100 a week. That gives you an idea of the magnitude”, confirms Tom.

“I firmly believe the incumbents now use AI (and machine learning and deep learning) primarily to save money. However, AI increasingly will be used to produce better and more affordable customer offerings by providing customized solutions according to individual needs. Let’s not forget—AI will give the ones who master it a clear and huge competitive edge”, says Spiros.

Does AI better suit traditional or emerging markets?

“For emerging markets, AI will be useful for coming up with new structural approaches and solutions in areas such as credit scoring and fraud and risk. For example, if you don’t have a lot of systems in place, AI will be able to create those structures quickly and robustly, all based on data.

“If you’re going to build a new risk or fraud prevention technology or identity platforms in emerging markets, you’re not going to do it based it on outdated methods, you’re going to use features such as biometrics and facial recognition for proving identity. So the potential for emerging markets to create new platforms to be leveraged from using AI and machine learning are huge”, says Brett King.

“For developed economies, it tends to be iterative, because you have embedded regulation in place. For example around identity, there’s the assumption that you have a web signature on a piece of paper, and to replace that with facial recognition is going to take a long take to get that through the regulatory cycle. But then you may see investments in AI in things like robo-advisors and chatbots, and the interface between customers, because those are the areas that you can get benefits with AI without issues on the regulatory side”, he adds.

“Once AI is mastered for a particular task, it can be applied regardless of markets. Of course, the traditional and rich markets will provide better and more customer data; so, therefore, the customized AI solutions will better work in traditional and developed markets”, Spiros comments.

“I think it will transforms most parts of our lives in all markets. I think we’re in the era of machine learning and AI and I think we’re only getting started”, adds Tom.

What are the biggest challenges facing AI and banking?

“That the data isn’t in one place and in a usable format”, says Tom.

“AI today is not going to solve the big banks main problems because of their complicated back office systems, so getting better technology in place will help without even implanting AI. Once you manage to get new systems in place that aggregate the data, then we’ll see the dramatic improvements.”

Brett adds: “Getting the right people is important. Silicon Valley is the epicentre of AI research. The best AI intelligence expertise aren’t necessary going to want and come work for a bank. Finding the right skills is key."

“The biggest challenges for AI are not the costs, computing power or technology, because the exponential computing advancements we see are faster than the legacy corporate and system structure can handle. I strongly believe trust, privacy and the fear that AI will kill more jobs than it will create are the biggest concerns the AI industry needs to address.” Spiros comments.

Expert commentary:

Tom Blomfield is CEO of digital-bank Monzo

Brett King is CEO and founder of Moven. As well as being an entrepreneur in the fintech industry, he is a best-selling author, radio host and keynote speaker

Spiros Margaris is a venture capitalist, thought leader and founder of Margaris Advisory. He is globally ranked the No. 1 Fintech and No. 2 Insurtech influencer by Onalytica