Microsoft's take: What’s next for AI in fintech: From experimentation to execution
AI is no longer future technology. It may still occupy the imaginations of sci-fi authors and screenwriters but for many businesses it’s a reality. It is, in fact, a reality for many consumers too, with AI hidden in everyday applications in ways they don’t expect, explains Daragh Morrissey, Director at Microsoft Worldwide Financial Services and one of our Fintech 5X5 Report contributors
There are, of course, high-profile use cases, such as evermore sophisticated virtual assistants, but these hide more everyday but no less effective applications. One of the most effective uses of AI is to find efficiencies in complex networks, such as supply chains and mobile networks. It can make fast decisions based on a lot of information. There are fields such as medicine where AI is quickly becoming indispensable.
Where are fintechs in all this? This is a tricky question given that fintech is a fuzzy term these days, ranging from scrappy startups a few years old to fintech unicorns and global banks. Zest Finance, for example, has a cutting-edge approach to AI. By defining fintechs as predominantly innovative new banks and high-growth start-ups, it’s clear they are making good use of AI but there is certainly more that could be done.
That’s not to say there aren’t challenges ahead. Fintechs are known for their boldness, but regulation is rightly causing pause for thought. As such there is a tendency for fintech’s to use AI to solve simpler problems rather than complex issues, closer to automation than realising the full potential of artificial intelligence.
Financial service providers should be asking where they could be applying AI now, and moving past the stage of pure experimentation. AI has a wealth of applications in financial services, so it may be probably faster to outline the areas where it won’t help than all the places where it will be embedded.
Shifting from experimentation to execution in the short to medium term is critical.
Fintechs may be at an advantage when it comes to AI, but this may be short lived. AI is likely to be adopted across the industry to solve many different problems, and those who make the most of the technology will be in a far better position than those who don’t. Addressing these potential speed bumps will be key.
People are the solution to the people problem
One of the biggest obstacles any business, not just financial services, will face when working with AI is the need for people who can make it work. AI is about as far from a “plug and play” technology as it is possible to get. Experts are needed at every step of the process of implementation—whether that’s defining the desired outcomes, training the AI, and tweaking algorithms to ensure the desired outcomes are being met.
There is fierce competition to hire and retain the right people. Building a team that can make a success of AI has the potential to be extremely difficult and expensive.
'The potential of AI is immense, but it’s key to build in ethics as soon as possible to avoid problems.'
To close the skills gap, Microsoft is offering new AI capabilities that make it much easier and faster to build AI applications for people without a data science background.
Does that mean fintechs that lack the deep pockets of bigger businesses may are at a disadvantage? Not necessarily. A culture of innovation may be more attractive than a bigger salary. People want to think that they are making a difference, and joining an agile fintech may hold more appeal than working for an incumbent.‘Your AI team may already be in your company, you just need to find it.
One solution to the people problem is to make the best use of the staff you already have. Microsoft is working hard to make the appropriate training available so that any business can make the most of AI. Training can democratise access to AI, and make sure it’s not just those who can throw money at the problem who will thrive. Your AI team may already be in your company, you just need to find it.
The need to build in ethics
There was a promise once that computers would remove human biases and make perfect decisions. But it’s clear now that AI can not only reflect human biases, but can amplify them. Any business that implements AI needs to recognise this issue and address it from the start. It’s far more difficult to address this issue in retrospect.
The fast decisions made possible by AI have many potential applications in financial services. Two of the biggest are credit decisions and detecting financial crime.
‘Your AI team may already be in your company, you just need to find it.’
Providers can let their biases get in the way of assessing risk, letting caution get in the way of potential revenue. AI is far better at assessing risk than blunt tools such as credit scores, which means providers can serve customers where they would previously be limited. Part of that better modelling of risk is eliminating biases that would leave certain groups at greater risk of financial exclusion.
AI also has the potential to save millions by attacking financial crime. The rules around AML and KYC go so far, but all it takes is someone savvy to what the rules are to find a way around them. Similarly, rules that protect against fraud are just that—rules. Rules can be broken and subverted whereas people can see shady business going on. But there simply aren’t enough people to find every suspect transaction.
AI can help every financial services provider to avoid fines, protect businesses from fraud, and drive inclusivity. The potential is immense, but it’s key to build in ethics as soon as possible to avoid problems.
Fintechs have the potential to keep leading on AI
Some incumbents have admitted to being scared by the potential of fintechs. And they should be. Smaller companies have the advantage when it comes building sophisticated AI. Without legacy technology to deal with, fintechs simply don’t have to revolutionise how they work to start implementing AI.
Some incumbents are working hard to narrow the gap, and some are already there. Fintechs may need to start thinking outside of their niche. Their immediate success has often been on doing one thing very well—how could AI apply to what they do? Should they be looking to partner with other fintechs on AI partnerships? If they lack the in-house skills, how do they solve this issue? By throwing money at it, or by building the capability in-house?
Whatever the question, fintechs should find answers quickly. Their advantage won’t last forever.