For the financial sector, the conversation has shifted from business agility to business resilience. What was once a battle of speed and first-to-market has now become a war on weathering the storm during highly volatile operational or security-related situations.
When these events occur, decision-makers within financial organisations need to analyse every scenario and act quickly – using data-extracted insights to guide the way.
In this piece, InterSystems’ Director of Product Management Jeff Fried explains how machine learning can play an integral role in building resilience within the financial sector, using real-time examples such as automated trading, fraud detection, and customer experience (CX) initiatives that utilise modern data management and analytics capabilities to make it through the most strenuous business challenges with almost no effect on the operations of the organisation itself.
Before the pandemic, increasing business agility was of utmost importance for organisations across industries. But what was once a battle of speed and first-to-market has now expanded its focus to include business resilience. Recent events have underscored the need to weather the storm during volatile operational or security-related situations. When these events occur, decision-makers within these financial organisations are forced to analyse multiple scenarios and act quickly, using data-driven insights to guide the way.
Financial resilience – the ability to withstand and recover from temporary financial hardship and disruptions – is not just an ability to react to conditions. It goes much deeper and requires building adaptive processes and maintaining customer relationships on a whole new level. Leveraging technology and data to gain insights and inform decision-making has become more important than ever before – especially within this highly dynamic and competitive landscape. As changes occur, financial organisations need to be able to react quickly – sometimes even before the change. Their ability to do this relies on healthy data and the use of analytics to extract insights out of that data rapidly.
Machine Learning’s Resilient Role
When implemented properly, machine learning (ML) can play an integral role in increasing business resilience for those within the financial sector. Machine Learning is intrinsically adaptive; it can adjust to changes in data, uncover the most important factors that drive outputs, and support a broad range of “what if” modeling.
By integrating ML into systems such as risk management, automated trading, fraud detection, customer service, loan underwriting, and marketing, business leaders will be able to make it through some of their organisation’s most strenuous challenges while minimising the impact on the operations of the organisation itself or its bottom line.
There are plenty of proven applications for ML in financial services, yet the number of actual production implementations of ML in financial services is still relatively low. In fact, financial services trails many other industries in both ML adoption and benefit. Why is such a quantitative industry lagging in machine learning? Among other factors, data challenges are the commonly cited issues among financial services firms struggling with implementing and deploying ML.
Gaining operational benefits from ML hinges on access to large and diverse data sets. However, attaining sufficient amounts of quality data continues to be a challenge.
Creation, capture, and consumption of data went up by a whopping 5000% between 2010 and 2020, and is projected to grow even faster in this decade. Data quality issues are brought to the forefront in this climate across all industries – poor data quality costs the US economy up to $3.1 trillion yearly – and financial services is no exception. In financial services, there is also a particular premium on tapping data in real-time, making things even more challenging. Organisations must grapple with these immense amounts of data and learn how to manage it intelligently and quickly.
Overcoming the Data Challenges
The way in which data is viewed by organisations has evolved, with data moving from being a by-product of business processes to one of its most valuable assets. But for many financial services firms, overly complex infrastructures that rely on a disjointed set of data management technologies are leaving them unable to leverage data fast enough – and in ways that can drive their ML initiatives forward.
These challenges have brought the need for next-generation approaches to data management to the forefront. Leading financial institutions, including Citi, Goldman Sachs, and JPMorgan are leveraging enterprise data fabrics to process, transform, secure, and orchestrate data from multiple, disparate sources in real-time, with full governance and a simplified data architecture to power their machine learning initiatives.
A well-governed, high-performance data fabric provides many benefits and can address many of the challenges in ML adoption. Enterprise data fabrics can access and harmonise data from disparate sources. In this kind of environment, machine learning is not simply hostage to data quality issues – it can be applied to assess and improve data quality.
Many of the roadblocks to the adoption of ML in financial services are being addressed with next generation data platforms that not only provide the core for developing enterprise data fabric architectures but provide high performance and low latency to handle large volumes of data in real-time. The most advanced technologies incorporate embedded analytics for developing and deploying machine learning models without moving the data. Unifying MLOps and DataOps under one data fabric makes real-time ML practical. Resilience is fully realised for real-time trading, customer personalisation, fraud detection, and a wealth of other applications.
Beyond data challenges, financial services firms also report a shortage of skilled data scientists. Solving the data challenge helps with this as data scientists spend an inordinate amount of time in data wrangling and strong data management with modern data fabrics can free up some of their time. Another advance is the advent of AutoML, which automates many of the mundane tasks of data science. Data fabrics that incorporate AutoML can free up data scientists to do deeper work and also bring the ability to develop machine learning models to a broader set of users.
Further volatility is ahead of us before we emerge from the effects of the pandemic, but with the use of modern data management, analytics capabilities and ML, financial organisations can reap the benefits of greater resilience and make it through to the other side with greater stability.