How AI can help banks survive in the zero interest rate world

A lot of countries enter 2021 with zero interest rates. According to CME FedWatch Tool calculations, the market does not expect any change here until the end of this year. The US has adopted an extremely expansionary monetary policy since March 2020. This is the response of the US central bank to the financial meltdown caused by the lockdown of the economy due to the Covid-19 pandemic. Interest rates in the USA have reached 0 to 0.25%, similar to the 2008-2015 period during the last crisis. The European Central Bank also eased their monetary policy in H1 2020.  

What do zero interest rates mean for borrowers? For customers with loans, lower interest rates mean lower loan instalments. Unfortunately, credit seekers have to accept higher "non-interest" costs, such as insurance and commissions. In times of crisis, banks run more risk when granting credit, they apply a very conservative policy, which makes it more difficult to prove creditworthiness.

What do zero interest rates mean for savers? Low interest rates on deposits not only fail to increase capital, but do not even provide protection against inflation. With additional savings account fees or taxes imposed by the government, it turns out that the saver often must pay extra. Zero interest rates are intended inter alia to increase the propensity to invest in riskier assets. Reallocation of savings towards assets that can provide a positive real rate of return requires accepting more risk. With employment and financial perspectives uncertain, the willingness to invest savings more aggressively varies widely.

Existential questions

How do you retain your current bank customers in the era of zero interest rates? How not to burden your customers with new costs? How to encourage new customers to choose our bank products? How to make the products offered by the bank more competitive? How to differentiating your offer in order to emerge from the crisis unscathed and build a strong position in the market? How to reduce costs without reducing staff costs and employment? There is only one answer to all these questions: New technologies.

The pandemic that has been going on for a year has accelerated technological progress and shown that the trend will be faster than expected in the coming years. Digitisation and automation of processes has been forced by, among other things, remote working on a larger scale than before and shutting down bank branches for fear of the virus spreading. New technologies are driving the transformation of banks. A prime example of that is Goldman Sachs Investment Bank which undergoing a transformation as well as introducing new digital products and services, reported US$ 10.78 billion revenues in Q3 2020, a performance that was higher than a year before and that exceeded the Wall Street analysts' estimates.

New technologies reduce both fixed and variable costs for banks and will thus help them survive a zero interest rate environment. Key processes automation also reduces the risk of high financial fines resulting from human errors. The risk of loss of reputation and credibility of the bank is then also reduced. This includes penalties for failure to detect money laundering. In the AML (Anti-Money Laundering) area, there is plenty of room for cost optimisation and quality improvement. One solution is to prioritise the reports coming from the rule system through the use of AI Risk Ranking solutions, which, based on artificial intelligence and supervised learning, develop a ranking of suspicious transactions according to their risk level. Faster addressing and more thorough investigation of riskier alerts allows the suspected money laundering to be reported more quickly and ultimately allows to avoid regulatory penalties. Prioritising also makes it possible to apply simplified procedures to alerts at the bottom of the ranking. Such alerts can also be subject to an auto-closed procedure. The use of Explainable Artificial Intelligence (XAI) will allow the AML analyst to focus on those elements of the transaction that affected the alerts' ranking.

RPA how does it work for AML, KYC and the others

Another tool facilitating the AML analyst work is a visualisation of financial flows between clients. The artificial intelligence makes it possible to analyse relationships in the data sets so large that humans cannot process them. As a result, the analysts are able to identify linkages between customers that they would otherwise miss.

Machine learning models often require a lot of computing power, but they do not need full power all the time. This is where cloud solutions show their strength, as the payment can vary depending on the time the resources are used, rather than, as in on-premises solutions, simply equal to the purchase price of the resources. This applies not only to the artificial intelligence solutions. Cloud solutions provide attractiveness and high availability while reducing costs.

Thanks to the use of robots, data can be promptly collected from both the bank's central systems and external sources. Basic as well as advanced analysis of the collected data can also be performed. Unsupervised machine learning allows for detection of anomalies in the data. All these operations will allow the bank employees to just process the collected information instead of searching for it in many sources. The employee's skills will be used for data analysis rather than data collection. Process automation (RPA) can be used for both AML, KYC (Know Your Customer) and new customer on-boarding.

Wealth Management is another area where AI can help to significantly increase the attractiveness of banking products. Predicting the expected rate of return for individual portfolios, taking into account external factors such as inflation, interest rates, media reports or sentiment analysis, among others, increases the attractiveness and competitiveness of banking products.

The key way for banks to increase revenues and reduce costs may be through process optimisation and automation. Solutions based on robotics, artificial intelligence and cloud solutions can effectively reduce these costs and increase the attractiveness and availability of banking products. 

Marzena Andrzejczak, Head of R&D, Comarch