Monday, 24 September 2018

Why Corporate Banks should be embracing AI and Machine Learning

With Fintechs and other challengers threatening to pick off the most profitable parts of Corporate Banking such as international money transfers and the provision of FX services, never has there been a more important time for banks to invest in fundamentally improving the ways that they serve their corporate customers if they are to retain and grow their share of their customers banking business and importantly delivering this profitably.

Unlike in the Retail Banking industry, where most customers use only one institution for their banking services, Corporate Bankers have always had to operate on the basis that their customers will have relationships with several institutions and therefore they have to compete for share of wallet.

There are two strategic questions that Corporate Banks need to address.

What is it that corporate treasurers (the principle owners of the relationships with the banks) want and what will incentivise them to increase the proportion of their banking business that they give to one institution over another?

A lot of what is driving corporate treasurers’ expectations today is coming from their experience as consumers. Given their experience of looking for a product on Amazon, choosing, ordering and paying for it and receiving their order the same or next day, their expectation of what a customer experience should be has been significantly raised. When even in Retail Financial Services where it is simple and cheap to make a foreign currency transaction using a Fintech such as TransferWise, it raises the inevitable question of why does it have to be so difficult to do the same in the corporate world? Even when TransferWise can’t make the payment instantly the consumer has complete transparency on where the transaction is in the process and importantly in real time. The corporate treasurer is looking for the same level of transparency and ability to self-serve when they engage with their banks. Having to phone their bank to find out the status of a payment is no longer acceptable.

Speed of execution is another expectation that has been changed by the treasurer’s consumer experience. They expect their banks to make decisions quickly and for transactions to be executed faster than they are today.

A frictionless experience in sharing data between the bank and the corporate is increasingly being demanded. Traditionally one of the reasons that corporates rarely change their banks is because the on-boarding process by banks takes a long time, is error prone, highly bureaucratic and every bank has its own process requiring slightly different information. If a bank can offer on-boarding that is frictionless, where the bank does most of the work and where the time to on-board is dramatically reduced then the positive impact on that bank’s share of the corporates banking business will be huge. Introducing a standardised approach to switching (where every bank asks for the same set of information and not asking for what they already know about the customer), as has been introduced in several countries in the retail banking industry, should be introduced for the corporate banking industry. If this was put in place there would be a dramatic shift in the number of corporates changing their banks. It is understandable why the incumbent banks don’t want to do this for fear of losing customers. However, those who do, and do it well will significantly benefit. If they don’t do it then one or more challenger banks will and will pick off the most profitable parts of their corporate business.

For the corporate treasury teams too much time is spent reconciling the cash accounts in their General Ledger with the bank accounts that they have with their banks. Much as open banking is promoting the idea of consumers having a single place where they can see all their accounts, regardless of which bank is the provider, so too Corporate treasurers do not want to have to visit a different portal for each of their banks but rather have one place where they can see all of their bank accounts. Simplification of that whole process so that there is a simple matching of Ledger cash accounts with bank accounts through the use of a virtual accounts solution allows the treasury team to focus on the important decisions about cash management. The bank that can offer this to their corporate customers will win a greater portion of their cash management and other banking business.

A frustration for the corporate treasurer is that their relationship manager often does not have a total view of the corporate’s relationship with the bank. Most banks are still organised around product divisions and it is left to the corporate treasurer to navigate around the bank’s organisation or worse still fend off multiple sales people from the bank trying to sell competing or overlapping products from the same bank.

The corporate bank customer’s requirements have evolved but are fundamentally straightforward and reasonable.

What role does Artificial Intelligence and Machine Learning play in delivering the Corporate Banking customer’s requirements?

Much as young children have grown up with the expectation that every device is touch sensitive and there is an increasing acceptance of Alexa and other voice-enabled devices, it won’t be long before a bank (or more likely a non-bank such as Amazon) will offer corporate customers a banking proposition where Artificial Intelligence and Machine Learning will simply and seamlessly be built into all business processes.

There is already evidence of it beginning to be used across the whole lifecycle of banking business processes. At the front end the use of Machine Learning to display the help pages in the order that they are most frequently requested, encouraging self-service by customers rather than them having to phone for assistance. In the back office it is beginning to be seen to be used for fraud and money laundering detection along with payment instruction repair.

Due to the difficulties of switching banks (as mentioned above), Corporate Banking customers have low levels of churn. However, what they do exercise is the ability to flex the share of banking business that they choose to give to individual banks. Identifying the leading indicators that a bank is becoming less favoured by a corporate customer is a task highly suited to Machine Learning. The key characteristics that lends to this being solvable using Machine Learning are the large quantities of structured (e.g. transactions) and unstructured data (e.g. social media, emails, phone calls) from a large cohort of customers. Looking back at common events that occurred before customers significantly reduced the share of their banking business with a bank should help to build an understanding of the leading indicators of business attrition. With significant returns if this potential loss of share of wallet is addressed prior to it occurring this makes it an ideal case for using Machine Learning.

The recent uncovering of large scale money laundering being enabled by a number of banks such as Danske Bank, Credit Suisse and HSBC and the subsequent consequences, both financially and reputationally, for the banks involved could have been identified earlier had Machine Learning technology have been applied to the problem. Machine Learning is particularly appropriate to this type of dynamic problem where the money launderers adapt their techniques and approaches to avoid detection and the system to identify and respond quickly to these changes.

Understandably one of the most frustrating experiences for corporate customers is when payments made are returned by the bank due to clerical errors such as incorrect IBANs, payee names or account numbers being submitted. Increasingly banks are turning to Machine Learning to fix these issues and allow the payments to go through without having to be returned to the customer. This is because of the increased IT ability to handle fuzzy data for instance where there could be names spelt incorrectly or digits transposed. Given the high volumes of transactions and the varying nature of the errors Machine Learning is far more productive at addressing this than manual intervention.
The changing demands of corporate customers, the increasing competition for the most profitable segments of banking business and the increasing cost efficiency of IT processing means that this is an ideal time for Corporate Banks to apply the power of Artificial Intelligence and Machine Learning to deliver a far better experience to their customers in a more profitable way.

Tuesday, 20 March 2018

Will GDPR inhibit or enable Open Banking?

Image result for "customer apathy"
GDPR is not the biggest threat to open banking, customer apathy is a far greater one. Banks and Fintechs have been pouring money into getting ready for open-banking, creating open APIs and new services and offerings for customers. However if there is one lesson that the seven-day switching service has taught us it is that the majority of customers are simply not interested in banking and see all banks as the same. Most customers would like to spend the absolute minimum amount of time thinking about their finances and see banking as a means to an end not the end itself. The volumes for seven day switching have been disappointing with an average of 75,228 per month in 2017. The expectation that there would be a mass move away from existing primary current account providers has not happened.

Even when customers have switched, it hasn’t been to either the neobanks (Monzo, Atom, Starling) or the challengers (Metro Bank, Clydesdale Bank, Yorkshire Bank, etc). With the exception of Nationwide Building Society, the net beneficiaries have been the large, global banks – First Direct (HSBC), Santander, TSB (Sabadell) and Halifax (Lloyds Banking Group). The neobanks are becoming secondary banks for the majority of their customers not the customer’s primary bank.

If seven-day switching hasn’t got customers excited about banking, will the offer of open banking be enough to get customers spending more with their existing banks or switching their primary banking relationship away from their current provider? If so, which banks are likely to be the winners?

Open Banking is intended to create more competition in the banking industry and to encourage better services and more innovation to improve the customer banking experience.

One way of improving the experience is  to provide a single place where a customer can see all of their banks accounts regardless of which bank provides them. This is not a new idea. Yodlee, the best known player in the aggregator market, has been around for over 17 years providing services to over 1,000 financial institutions and fintech providers. Account aggregation, which sounds like a good idea, has not taken off in the mass market. Apart from the customer apathy described above, the screen-scraping technique deployed by many aggregator tools involves the customer breaking the terms and conditions that they had agreed with their banks. This is where GDPR, Open Banking and PSD2 (which use open APIs) jointly provide a regulatory framework to give consumers the knowledge, should they wish to take up such services, that they are legally protected.

GDPR is about putting consumers back in control of how their data is used. GDPR from a customer’s perspective is a pre-requisite for open banking as it will give them the confidence that their personal data will only be used for the specific purposes that they have had to explicitly agree to when signing up for the service.

Account aggregation is not the only new service that banks, fintechs and non-banks are beginning to offer to customers. Real-time spending analysis, the ability to split restaurant bills and lower cost foreign transactions are among the services that both existing and neo-banks are offering.

A question that the banks must answer is whether the current open banking offerings are providing an experience that is sufficiently differentiated from the competition that it will make customers actively switch to them.

Neobanks being built using cloud first, modern technologies have advantages in both complying with GDPR and offering new services as a result of open banking. They have had been able to build from the start a single view of the customer in real time using open APIs and microservices. However, they lack scale in terms of both the numbers of customers and the depth of resources.

The existing big five banks have all the advantages of the size of their customer base and IT budgets. However they are hampered by the complexity of the legacy infrastructures, customer data is spread across multiple legacy systems designed for batch-processing makes building a real-time view of a customer’s relationship with the bank a significant challenge. This is why a number of the major banks have either elected to work with Fintech firms to help them address this or have designed new digital banks using modern technology.

For banks and non-banks (since the legislation was drawn up to encourage challengers from other sectors such as telcos, retailers and fintechs) GDPR increases the potential financial and reputational risks of entering the open banking market. While most people know little about the detail of GDPR almost everyone seems to know about the fines of up to 4% of global revenues for a breach of the regulation. The regulation goes live on 25th May 2018 and no organisation knows how strictly it will be enforced and certainly don’t want to be the test case for the first fines.

Given the risk of fines and the cost of meeting regulation, the revenue upside of entering the open banking need to be significant. Providing an aggregator service or a breakdown of expenditure in real time are good customer experiences but don’t directly bring in additional revenue as the neobanks are finding. Open banking is of course about more than just providing aggregation and PFM (Personal Financial Management) and the revenue growth is forecast to come from the provision of additional financial and non-financial services. All of the neobanks have realised that offering current accounts alone is not a profitable business. To be successful they need to be able to offer other services and are positioning themselves as marketplaces. One of the most successful organisations operating as a marketplace has been Moneysupermarket, but even they are finding that competition is driving down their margins and the barriers to entry (helped by the intervention of regulators) has significantly impacted their profitability.

A key criteria to be a successful marketplace is to have scale – amazon, ebay and Ariba (in the b2b world) demonstrate this. As open banking becomes a reality then the winners will also be the ones that have the scale. For the moment that advantage lies with the incumbent banks.

The success of open banking will neither be enabled or inhibited by GDPR. The success of open banking in the retail segment will be measured by the level of switching activity significantly rising.  This will only happen by providing an offering that so engages the customer that it overcomes the disinterest that most customers have about banking.