The key technologies driving the future of finance

With a presence in over 100 markets, more than US$2.6 trillion (S$3.5 trillion) client assets under management and more than 250,000 employees, JPMorgan Chase is a banking behemoth. Yet this financial institution, with a history dating back more than 200 years, is also trying to adopt a start-up’s nimbleness.

It invests US$12 billion a year in tech, some of which has been invested in artificial intelligence (AI) and machine learning (ML) – two fields that offer revolutionary new opportunities. For instance, JPMorgan Chase has rolled out an AI-powered virtual assistant that helps corporate clients manage their global accounts by making insightful recommendations based on intelligent analysis of the client’s behaviour.

Its research entity, J.P. Morgan Markets, is a prolific platform that produces over 10,000 pieces of research a year. However, clients were not always aware that this wealth of material existed. ML techniques recently solved this issue. When clients log onto the platform today, they are provided relevant research that has been personalised to meet their needs.

JPMorgan Chase is just one example of how more firms are turning to AI and ML to overcome a variety of business challenges.

Other platforms and solutions like HSBC Evolve – a customisable foreign exchange platform – and HSBC MyDeal – a digital platform solution that optimises capital markets transaction experiences – also illustrate how AI and ML are making their mark in fintech.

Companies that are prepared to leverage AI and ML can use these technologies as cornerstones to innovate, advance and grow beyond competitors – a tenet that is expanded upon in a new report from LSEG Labs titled The Defining Moment for Data Scientists.

The report is based on LSEG Labs’ annual independent survey of 482 data leaders and practitioners in financial services across the globe.

MAKING THE BIG LEAP

The JPMorgan Chase and HSBC examples illustrate one area of interest in the report: AI and ML can greatly enhance the customer experience, especially when design thinking – the process of trying to understand how users interact with a product or service – is incorporated into such projects.

In general, companies expressed enthusiasm for improving and personalising the customer experience through AI or ML, but had reservations over possible regulatory issues. While a majority expressed confidence about their progress in deploying AI or ML, they also found challenges in budgeting as well as nailing end-to-end processes that could deliver on the promise of AI and ML. 

MANAGING AND CONNECTING DATA

Another trend highlighted in the report was the increasing integration of unstructured data into models, with fewer companies reporting that they worked exclusively with structured data.

Increasingly, companies are supplementing financial data with alternative data, as well as spending more time annotating data. In the last two years, alternative data has received a boost from COVID-19, as the pandemic’s disruptive impact on both global and domestic supply chains has resulted in rising interest in tracking data such as shipping and sensor data and satellite imagery.

However, poor data quality and lack of availability still pose significant barriers. The former challenge is partially resolved by the use of synthetic data, which is wielded by a majority of companies surveyed. When it comes to managing data assets, though, many reported a lack of processes that resulted in numerous issues in connecting and aligning different data sets.

PARTNERING WITH EXTERNAL VENDORS

More financial services firms are migrating to cloud infrastructure and outsourcing AI and ML services, with a significant number seeking out third-party vendors that can integrate with internal systems.

Such vendors can offer wide-reaching, powerful tools that nevertheless are easily adapted to a client’s specific needs. For example, in a recent interview, Mr Geoff Horrell, global head of innovation at the London Stock Exchange Group (LSEG), highlighted the customisable aspect of LSEG Labs’ AI and neural network-powered pre-trade market impact analysis tool. The tool relies on six months’ worth of historical tick data from the S&P 500 and Russell 1000 indexes, combined with data from other markets. “Customers can see the analytics and add their own features to those analytics,” he said. “It’s a more open way of delivering the analytics.”

In addition, natural language processing (NLP) has become a core focus area, with many firms preferring to rely on application programming interfaces from the largest cloud providers. Supervised learning has also fallen out of favour, with reinforcement and unsupervised learning growing in adoption while deep learning continues to dominate.

RISING EXPECTATIONS

As the world begins to come to terms with the pandemic, the uncertainty of 2020’s hiring expectations has given way to a more bullish outlook – more respondents in 2021 expected to see an uptick in the number of data scientists employed in their companies.

The role of data scientist is also evolving into a more strategic one. The job scope has been transformed with greater responsibilities, such as building and deploying models, selling the business case and influencing internal strategy. More specialist roles, such as data engineers and data architects, are in demand, while employees with domain expertise are eager to improve their coding skills.

A MODEL TO GOVERN BY

Along with the maturation of AI and ML comes the need for model governance frameworks. As data science teams expand, processes improve and model inventories grow, model governance is essential to ensure model quality and reduced costs.

Those surveyed predicted that as AI and ML deployment moves from experimentation to production, model governance will be increasingly implemented for statistical robustness, rather than to mitigate regulatory risk. But in the long run, some respondents expected ethical concerns to be a bigger driver of model governance implementation, citing bias and the issue of explainability as growing challenges to be resolved.

Stakeholder engagement and the communication of benefits are also expected to take higher precedence in order to raise stakeholder trust in AI and ML – a necessary move, as the technologies continue to disrupt sectors globally and reshape the playing field for most, if not all, industries.

Download the report from LSEG Labs for more insights.