Marcin Kacperczyk is a Professor of Finance at Imperial College Business School, and teaches on the AI & Machine Learning in Financial Services virtual Executive Education programme
Our workplaces and industries are becoming increasingly more digitally driven, and the Financial Services sector is no exception. According to a report compiled and published by the Economist Intelligence Unit at the end of 2020, 54% of financial services organizations with more than 5,000 employees have already adopted AI technology, and smaller financial services providers are not only taking note, but following suit. A Bank of England survey has reported that financial services firms expect to see significant growth in AI and Machine Learning technologies over the next three years.
And it’s not hard to see why. Such technologies not only speed up and increase capacity for the day-to-day basic tasks such as transferring and moving money or making payments, they boost convenience for customers and companies alike. This latter point has been particularly important over the past 18 months as Covid-19 has dramatically altered the ways in which traditional business is conducted. A switch to digital and automated services within the financial services sector has reduced the need for face-to-face meetings, made services available to customers 24-7 and has provided greater visibility and autonomy to customers at a time when trust has been paramount.
But these technologies have even smarter and more nuanced uses which, if deployed correctly, stand to provide huge returns. For example, by gathering customer and industry data and identifying algorithms within it, such technology can improve decision-making and reduce the potential for human oversight and error – not to mention reach conclusions in a quicker, more accurate manner. Aside of such capabilities leading to increased productivity which, in turn, boosts company and industry performance and profitability, the insights they can provide can give companies the inside track on how the wider industry is operating and how they can use this to their advantage.
However, investing in, and making effective use of such technologies is not as straightforward as it seems – and can be the cause of significant losses if not deployed correctly. Alongside those industry reports predicting huge levels of digital take-up come the less-optimistic reports from the likes of Gartner, a global research and advisory firm, which predicts that up to 85% of AI projects will result in flawed outcomes due to bias and mistakes not only in the data they collect but also due to the humans in charge of managing them.
We cannot slow or oppose the impact that technologies such as artificial intelligence and machine learning have had on the world, so how can financial services professionals help ensure that they can keep pace and be able to work effectively in an increasingly smart landscape?
The key to making good decisions about AI and Machine Learning – and avoiding expensive failures – is, unsurprisingly, being able to understand it. Of course, firms can invest in basic training for staff by enrolling on education programme at business schools specifically designed to upskill financial services professionals. But aside from this, there are a number of other simple steps professionals in the sector can take to help protect against bad investments or poor decision-making when it comes to bringing such technologies in-house.
Through my own research, and in working with financial services executives on the AI & Machine Learning in Financial Services Executive Education programme at Imperial College Business School, I have identified the common missteps most financial services professionals take when it comes to AI and machine learning deployment, and from this have devised five simple, yet essential, lessons for them to learn.
- AI and Machine Learning are different
It might sound obvious from the fact that the two have different names, but it’s easily to confuse their purpose as they provide very similar functions. However, the subtle differences between the two are not to be ignored as they operate in wholly different ways and have different requirements for their success. Being able to make the distinction between the two is vital for financial services professionals in order to better decide where and how to use them – or even if they need to be used at all.
AI, very simply, creates programmes which mimic then enhance human processes. However, it has its limitations. Current AI technology is routine-based – learning from common data patterns and making decisions based on the typical actions and results of similar scenarios it has previously recorded. Because of this it cannot forward think or innovate in the same way a human can. It is limited to only making decisions and providing results based on probability from historic data. The sci-fi perception of self-aware, self-thinking AI bots are no more than fiction so far.
But Machine Learning might be the closest thing we have to autonomous AI so far. It exists as a sub-category of AI’s capabilities. It’s a technology designed and implemented with a specific purpose – to either fit models or identify patterns in the data it is provided with, without explicit programming or needing human intervention. A good example of Machine Learning in action is how a search engine might rank page results for the words and phrases that users type into them.
- Train your Algorithm
The likes of Google may have AI “super brains” that can outsmart world champion human chess players, but it is misleading to think that these machines have secured such results under their own intelligence. These super-computers are the result of hours upon hours of human intellect, filling them full of data and teaching them to follow and identify certain traits within it. They cannot outsmart humans on their own, instead they’re making judgements of probability, having been fed all the possible information there is to know about the topic at hand.
It’s a fantastic example of how, with the right lever of detail and understanding, Machine Learning can be applied to great success in scenarios where there are clear boundaries and a finite set of outcomes (like the rules of Chess). This sort of structure is where Machine Learning can really thrive.
Apply this thinking to the finance sector – if trained effectively, such Machine Learning algorithms could tidy-up messy, complex data sets with a slimmer margin for error, and suggest best routes forward. For example, Machine Learning can help analysts distil hundreds of potential indicators of future investment returns into a few, more robust, measures – something much more manageable for human professionals to work with. Which leads nicely into my next lesson
- Machines cannot beat human rationale
These technologies excel at identifying relationships and patterns within a set collection of data; however, they are incapable of doing so in less structured environments. And this is where effective human decision making and rationale must come into play. It is not just unsuitable but also potentially dangerous to attempt to use a tool such as Machine Learning to solve problems where discretion or special consideration may need to be made, or where ethics might be brought into question.
By its very nature, the finance sector places great stock in human instinct and taking calculated risks. This is a skill that surpasses even the most intelligent of artificial services. Whilst an algorithm might well do a good job of identifying patterns in historic data to help provide the information on which to make a judgement call, its intellect is limited to past performance which, as we know is not always an accurate indicator of future potential. For this reason, humans are still very much setting the standard in the sector and will remain the driving force for some years to come.
- Good data is vital
Good data gets good results. For your Machine Learning application to be effective you need to have reliable data for it to work with. And that can be hard to come by. It is a misconception that the finance industry is awash with data. Whilst this may well be true for transactional areas such as payments, it is typically very hard to gather effective data in areas such as company performance. This makes it difficult for analysts producing quarterly reports to build a statistically robust machine learning model. In addition, in areas where there exists a grey area between how data is compiled or categorised it becomes difficult to then produce algorithms that are free of bias.
- Machine Learning can go beyond numbers
Despite the many learning hurdles, there are good opportunities to use Machine Learning to enhance financial decision making, and as the technology improves so too will the advantages we can gain from it.
For example, software is currently being developed to allow Machine Learning to go beyond numerical analysis and instead conduct accurate textual analysis too. This development will soon make it possible to analyse customer’s communications as well as their actions. Additionally, image analysis capabilities are also being developed which could, one day, prove to be a real asset to the financial sector by creating more timely and reliable data. For example, an analyst could use real-time satellite imagery to record the number of cranes being put up in a city and use this information to help measure the levels of construction activity over a given time-period, which could then help produce information in advance of industry surveys. A far-off prospect, but one absolutely within the realm of possibility.
And these developments mean it is vital for financial services professionals to keep on learning – and avoid investing in AI and machine learning because of the hype, but to instead gain a truly competitive advantage. Failing to understand how and where such technologies can best be applied will not only result in loss in the short term, but also result in a significant disadvantage as these technologies develop and the focus turns to not only what can be possible, but ethically what should be possible.
As Machine Learning’s ability to digest different types of data expands there are questions which need to be asked about how to ethically use data from other sectors, particularly when it comes to using this data as part of the decision-making process. Those that can face those discussions and master the technological application process effectively stand to win big.
Most crucially, for that development to happen, it is vital that practitioners take the steps now to learn the current state of the technology, where it works well and where it doesn’t. Master that and you can master all that is to come.
THE ACCELERATION TOWARDS A MOBILE FIRST ECONOMY
By Brad Hyett, CEO at phos
Over the last year, we have seen a big shift towards contactless payments. Fuelling this has of course been the coronavirus pandemic, which has made the public hesitant to handle cash due to the health concerns.
As multiple national lockdowns forced physical stores to close, and customers demanded easy, cash-free payment options, merchants had to quickly adapt. The result? An increased provision of pay and collect services.
In the UK alone, 83% of people use contactless payments according to data from the Office of National Statistics.
So it’s vital that merchants are equipped with the most efficient payment solutions, as the UK heads towards a mobile-first economy.
Proliferation of contactless payments
In 2020, 90% of UK card payments were contactless. This equates to an increase of 12% on the year prior, despite the total number of payments made falling by 11% from 2019 to 2020. Moreover, the affordability of smartphones has increased significantly over the last decade. And it’s estimated that 84% of UK adults now own one.
We’re Seeing merchants embrace more efficient and cost effective payment methods in response. While physical payment terminals are often too expensive for many small businesses, software point of sale, or SoftPoS, enables merchants to turn hardware that they already own – i.e. their mobile device – into a point of sale terminal.
With merchants increasingly adopting these innovative technologies, contactless payments will continue to gain popularity among the general public. In 2020, 13.7 million people in the UK either didn’t use cash at all or only used it to make a single purchase. That’s double the same figure from the previous year.
Changing consumer demand
Now more than ever, consumers are aware of how innovative payment solutions can add efficiency to their daily lives. As such, consumers now demand better payment services, including reduced queuing times, checkoutless stores, and bespoke loyalty schemes.
Businesses such as Mercedes offer an end-to-end digital car purchasing service, so customers can go through the whole car purchasing journey from the comfort of their own home. This includes car deliveries, financing, insurance and more.
Meanwhile, eCommerce giant Amazon has started trialling checkoutless ‘Go’ stores, speeding up the shopping experience by eliminating the queuing process altogether. The days of waiting for a table at a restaurant are also over, as more people have grown used to booking in advance.
Hence, it’s important that we empower small businesses to remain competitive and provide them with the payment solutions to meet customer demand.
The digital payments revolution isn’t slowing down anytime soon. By 2026, only 21 percent of transactions will be made using cash.
The US might have been slow out of the gate, but it’s starting to see increased adoption of mobile payments. In-store mobile payments grew by 29% in the States last year alone.
This growth was primarily fuelled by Gen Z-ers and millennials. Latest projections show that there will be 6 million new mobile wallet users by 2025, with millennials accounting for 4 million of this figure. These two generations, the former in particular, have grown up with mobile banking.
For most Gen Z-ers, their first foray into financial services was with a challenger bank like Starling or Monzo. These banks are able to offer online features such as ‘split the bill’, fee-free withdrawals abroad and much more to cater to the modern financial needs of the younger generation.
The Middle East experienced similarly sharp increases in contactless payments. From 2019 to 2020, there was a 200% growth in contactless transactions. This shift towards a mobile-first economy in the region was inevitable; the pandemic merely accelerated this shift. A recent study showed that 80% of people living in the Middle East planned to continue using contactless payments post-pandemic, with speed and security being the main draw.
The future is mobile
As parts of the world now start to come out of lockdown, there’s an openness to new solutions and a widespread acceptance of new technologies.
It is now a case of when, rather than if, we’ll see a permanent shift to cashless in the future. For businesses, embracing digital innovation will be key to remaining competitive and keeping pace with consumer demand in this fast-changing payments landscape.
HOW MERCHANTS CAN IMPROVE THE ONLINE PAYMENTS EXPERIENCE
By Alan Irwin, Senior Director of Product at Global Payments UK
The dramatic increase in online shopping over the past 18 months has encouraged many businesses to invest in developing their omnichannel shopping experiences. The reasons vary – some are keen to capitalise on the trend of older shoppers migrating towards ecommerce and some are trying to make up for loss of sales in brick-and-mortar stores during the pandemic. It is also true that many businesses are shifting their models to sell direct to consumers to avoid high marketplace fees and are therefore building their ecommerce channels for the first time.
The checkout experience is arguably the most important and delicate part of the ecommerce transaction, as it can make the difference between a happy customer likely to return, and a shopping cart abandoned out of frustration and confusion. A survey from March 2020 suggested that 88% of online shopping orders were abandoned, i.e. not converted into a purchase. A seamless, customer-centric online payment experience is therefore critically important in ensuring completed transactions. But with so many payment providers available, what should businesses be looking for when trying to keep friction to a minimum?
Keep clicks to a minimum
Less touchscreen interaction equals less abandonment. Adapting the payment page to fit any device and supporting popular mobile digital wallets like Google Pay ensures a seamless, stress- and hassle-free checkout experience for the customer and keeps clicks to a minimum. Friction can present itself in the most minor features – for example, when the customer is navigating the payment form, the appropriate keypad should be shown to the customer when required. It’s much easier to enter a card number using the dial pad instead of switching between QWERTY keypad layouts.
Simplifying online forms with autofill and tokenisation also significantly reduces friction at checkout and shortens necessary time taken. Ensuring checkout forms are tagged correctly for “autofill” is a great way to offer customers a single-click to input the payment, shipping, and billing data that they have stored in their browser profile. Similarly offering a guest checkout option will help convert customers who are in a hurry or looking for a one-off purchase. This can also be achieved by offering to store the payment details (called ‘tokenisation’) for express repeat and one-click purchases.
Make it easy to understand
A tailored payments approach can increase both domestic and international global sales. By offering a checkout experience in the customer’s language, the option to pay in their currency of choice, and use their preferred method of payment (whether it’s PayPal, Alipay or card), businesses can build loyalty quickly and put customers at ease. It is equally important for merchants to ensure they always display simple direction and information about next steps to instil confidence and prevent customer drop-off. The customer should be informed of what is happening at every stage in the process, for example, whether they will proceed to SCA (Secure Customer Authentication) next or go straight through to completion.
In addition, validating forms in real-time means merchants can highlight potential errors to the customer early on, and payment providers should provide this functionality. This could be an invalid expiry date, an incorrect digit in the card number or incorrect CVV number based on card type. When issues are only flagged at the end of the process, this forces the customer to go back through the steps to figure out the error. Real-time signposting of problems removes this potential friction and reduces the potential for a declined transaction.
Ensure seamless security
Merchants should work with a payment partner who offers the right blend of security and compliance management without it coming at a cost to the end-to-end checkout experience for the user. Instilling trust and security in your checkout flow while utilising the right solutions to drive seamless authentication flows will increase customer confidence and help prevent drop-off.
The greatest level of security and control comes from either utilising hosted payment fields that the
merchant can natively integrate into their checkout flow, or a hosted payment page where they can
manage the look and feel. Showcasing your brand on the checkout page with trust signals and logos also adds to building trust with the customer.
Staying ahead of regulations is also important. Secure Customer Authentication (SCA) will soon be mandatory in the UK for all eligible digital transactions, and this doesn’t have to be a friction-full process. Tools like Transaction Risk Analysis (TRA) and Exemption Optimisation Service (EOS) can quickly score transactions and drive exemptions where there is the right blend of transaction risk.
The devil is in the details
These three rules for successful ecommerce checkout experiences may seem straightforward, but it is important to apply them at a micro level. It can take only one minor point of friction to cause a customer to abandon their cart, and this will inevitably be replicated across other similar customers. It is critical to identify friction points early on and anticipate customer needs throughout the process. Discussing these points and any opportunities to improve customer checkout experience with your ecommerce team and payment provider is an important first step towards ensuring your entire shopping experience remains competitively seamless and loyalty is won. It may be that your payment provider cannot address them, in which case it could be time to move on in order to stay competitive.
FINTECH COMPANY PAYEN CHOOSES AQILLA FOR ITS LIMITLESS SCALABILITY AND SUPERIOR MULTI-CURRENCY FEATURES
Payen is a fast-growing FinTech company that provides gateway Payment and FX services to online merchants. Having launched in 2010,...
THE ACCELERATION TOWARDS A MOBILE FIRST ECONOMY
By Brad Hyett, CEO at phos Over the last year, we have seen a big shift towards contactless payments....
NEW RESEARCH REVEALS KEY ROLE OF KYC COMPLIANCE IN DRIVING CUSTOMER LOYALTY, ADVOCACY AND NEW BUSINESS
The impact of financial crime for institutions goes beyond crippling fines A piece of original research conducted by RegTech...
HOW MERCHANTS CAN IMPROVE THE ONLINE PAYMENTS EXPERIENCE
By Alan Irwin, Senior Director of Product at Global Payments UK The dramatic increase in online shopping over the...
JUMP-STARTING PROCUREMENT TRANSFORMATION WITH A CLEAR AND REALISTIC PLAN
by Alex Klein, COO at Efficio Consulting Following a period of ongoing economic uncertainty, business spend has risen high...
NAVIGATING FINANCIAL SERVICES IN 2021: LOW-CODE TO THE RESCUE
Nick Ford, Chief Technology Evangelist, Mendix Financial services are the poster child of great digital transformation: today, Britons can...
PAYSAFECARD AND NEO EXTEND THEIR SUCCESSFUL PARTNERSHIP
paysafecard, a market leader in eCash payment solutions, and NEO, one of the most successful FIFA teams in the world,...
WHY THE NORDICS WILL CONTINUE TO LEAD THE WAY IN DIGITAL PAYMENTS
Kriya Patel, CEO, Transact Payments While the recent introduction of PSD2 — the second iteration of the EU’s Payment...
COMBINED RISE OF M&A AND CYBER RISK CREATES STORMY SEAS FOR INVESTORS
UK organisations carrying out merger and acquisition (M&A) activities must improve pre-acquisition due diligence of software vulnerabilities By Philippe Thomas,...
PPRO CLAMPS DOWN ON FINANCIAL CRIME RISKS, PARTNERING WITH AND INVESTING IN AI-DRIVEN TRANSACTION MONITORING STARTUP SENTINELS
PPRO, the leading local payments infrastructure provider, has today announced a strategic partnership and minority investment in Sentinels, Europe’s leading transaction...
EMV® IN TRANSIT: WHY AND HOW?
Taoufik Sakhi, Smart Mobility Technical Advisory Director at Fime Today, contactless cards provide a fast and frictionless payment experience,...
INSTANDA ENTERS THE MIDDLE EASTERN MARKETPLACE
INSTANDA expands global footprint by working with new client, NewTechMe First product distributed in the Middle East Announcement signals INSTANDA’s understanding of NewTechMe’s vision to drive digital transformation in UAE...
RGU LEADS EUROPEAN INTER-REGIONAL NORTH SEA PARTNERSHIP TO HELP HOMEOWNERS IMPROVE ENERGY EFFICIENCY
NB: Image from left to right includes: Mike Bauermeister, Kishorn Insulations, Jamal Alabid, RGU, Amar Bennadji, RGU, Richard Laing, RGU,...
JUMIO APPOINTS JENNIFER N. HARRIS TO BOARD OF DIRECTORS
Addition of veteran CFO comes amid period of record growth and product expansion at Jumio Jumio, the leading provider...
WISE LAUNCHES ASSETS, YOUR WISE ACCOUNT INVESTED IN THE WORLD’S LARGEST COMPANIES
Assets offers current account flexibility, with the potential for investment returns Wise, the global technology company building the best way...
A CHECKLIST FOR RETRENCHMENT READINESS
By Shelley van der Westhuizen, head of financial well-being strategy & applied research at Alexander Forbes Your health may not...
EQUIDUCT LAUNCHES TRADING IN EXCHANGE TRADED FUNDS FOR RETAIL INVESTORS IN EUROPE
Equiduct will offer 436 ETFs and ETPs for trading through Apex Equiduct, the pan-European retail exchange, announced today that...
THE IMPORTANCE OF MANAGING DATA RISK IN THE FINANCE FUNCTION
Written by Steph Charbonneau, Senior Director of Product Strategy, Vera by HelpSystems CFOs and financial controllers play a pivotal role in how organisations evaluate and manage...
THE DEMAND FOR BETTER B2B PAYMENTS
By Brandon Spear, CEO, TreviPay Business-to-consumer (B2C) payments started adapting to digital processes when consumer shopping habits began shifting...
HOW TO BUY USDT AND AVOID THE HIGH VOLATILITY OF CRYPTO
Understanding and breaking down all the different types of crypto can feel like a huge task—there are so many variations...