While there are troves of financial data that have the potential to drive and inform critical business decisions, data remains siloed due to challenges associated with managing multiple data sources and aging infrastructure. Many financial institutions (FIs) face challenges – from limited storage and compute capacity to resource-intensive network deployments – that impede the ability to generate data-led insights powered by analytics.
Finextra spoke to James McGeehan, global capital markets lead - partners, Jessica Kowalski, global alliance lead financial services and Alvin Huang, capital markets specialist, at Amazon Web Services about how cloud is helping FIs upgrade their approach to financial data using artificial intelligence (AI) and machine learning (ML).
Laying the groundwork
Data is at the centre of digital transformation. FIs use it to optimise their business operations, enhance customer experience, bring new products and services to market, and gain competitive advantages. Capital markets firms, in particular, are aggregating traditional market data with alternative data, back-testing trading and investment strategies, delivering more granular regulatory reporting and more.
Sourcing financial data, acquiring it and delivering it to key systems, however, can be a complicated and expensive process for FIs. Requirements includes integrating multiple data sources and network providers as well as determining geographic delivery of data. Daily feeds from these vendors require heavy DevOps support and security protocols.
Additionally, FIs can’t share data using the internet like an average consumer. There are a multitude of complexities and regulations that determine how and where FIs can store and send financial data.
“The cloud directly addresses these challenges by simplifying access to data and data management as well as accelerating analytics. Kowalski highlights: “The cloud allows business to create, store and analyse data like never before so data workflows are deeply intermingled in the journey to the cloud. Most often there is executive commitment to be all-in the cloud and movement of various data assets to storage, but the devil is in the details. Mapping the course to weave together disparate data systems and third-party data sets in the cloud is a key requirement for AI and ML readiness.”
While FIs have been experimenting with AI and ML for years, the integration of these technologies into day-to-day operations was slowed due to a lack of in-house data science expertise and insufficient experience manipulating large data sets. Times have changed.
McGeehan explains that today, “FIs are increasingly investing in AI/ML, thanks in part to the availability of cost-effective, easy-to-use and scalable AI/ML cloud services. FIs are using these types of tools to enhance customer interactions through chatbots, improve surveillance, mine trading ideas from unstructured data and customise product offerings, among many other use cases.”
Kowalski has a similar view and focuses on these tangible examples of innovation. “Five to seven years ago, the focus was on the earlier stages of the financial data lifecycle like production and storage with firms concerned with discovering and migrating data assets from disconnected systems and data centres.
“Now that cloud adoption has gained momentum, they are paying more attention to deriving insights, predicting future events and optimising the flow of data between stakeholders. This naturally intersects with AI/ML. The two major areas of innovation I’ve seen are AI-powered chatbots in the consumer banking and investment space and algorithm/model training among the quant hedge funds.”
Symphony, for example, utilises chatbots to facilitate a broader sense of collaboration, changing how FIs share information. Specifically, getting the right people together with the right information in real time can mean more profitable decisions.
Cloud is a big driver in simplifying AI/ML, making the technologies more accessible across an enterprise. From developers and data scientists to researchers and business users, FIs are building intelligent applications more easily than before.
IHS Markit believes that having a single source of truth for business and operational data, especially where ML models are concerned is critical. A trusted record that can be traced through the lineage of governance takes data management to the next level.
This is one way to unlock innovation for new products and services. There are more according to McGeehan: “Financial services institutions can now rapidly test multiple data sources to challenge long-held investment processes to glean new insights. They are driving faster decisions, greater research depth, enhancing customer experiences, and promoting client retention and acquisition.”
Refinitiv, for instance, is finding new ways to generate investment ideas, including delivering customised sentiment analytics. These enable asset managers to make informed decisions more quickly and easily.
Kowalski adds: “The cloud has created an environment ripe for disruptive data innovation. Whether its currency trading, banking or credit swaps, the seamlessness of data integration, the accuracy of insights and the ability to share that information quickly is the heart of how firms gain a competitive edge.”
According to IDC, 40% of digital transformation initiatives in 2019 were supported by AI. In 2020, the expectation is that FIs will expand their use of AI and ML from front office applications to compliance and the middle office.
Huang predicts, “As regulations such as the Consolidated Audit Trail require firms to collect more data, we expect financial institutions to increasingly apply machine learning to enhance their markets surveillance logic.
Additionally, with the LIBOR set to expire at the end of 2021, we anticipate that financial institutions will accelerate their use of AI/ML in areas such as optical character recognition and natural language processing to identify and re-paper existing contracts that reference LIBOR."