Financial institutions possess petabytes of data that can provide insights to identify new sources of revenue, attract and retain customers, and reduce risk. Yet according to IDC research, 87% of chief data officers across the financial services industry cite difficulty with data availability, accessibility and quality within their organizations. Coupled with this, end users and customers have come to demand faster responses and at their convenience, turning what might be seen as a data problem into a business one. Those who innovate will win over customers from those who hesitate.

To fully capitalize on this amount of data and unlock its benefits, organizations are increasing their adoption of machine learning (ML) to extract insights from large volumes of structured data as well as data from voice recordings and text.

Human-centric automation

One popular application of ML in financial services is the use of speech analytics to optimize call center processes and improve customer experiences. Traditionally, call centers will store all of their customer call conversations, but review and analyze only a small percentage of those calls due to the time constraints. Organizations that leverage ML are able to analyze their full set of voice data in real time.

John Hancock, one of the largest providers of life insurance and financial products in the United States, leverages ML to provide sentiment scores, real-time insights and recommendations to its call center agents so they can better serve their customers. For example, by leveraging ML capabilities such as voice transcription and natural language processing (NLP), the application would be able to understand why the customer was calling and automatically prompt the call center agent with instructions, answers and links relevant to the issue, decreasing the time it takes to resolve the customer’s issue.

Predicting product and service demand

Financial institutions are also harnessing ML for predictive analytics, extracting insights from their existing data to get ahead of potential issues and recommending relevant products or services before they are even requested.

For instance, Broadridge, a leader in shareholder proxy voting services, extracts millions of data points from more than 300,000 shareholder meetings going back over 10 years. Broadridge now leverages ML to automate the extraction of data points from SEC regulatory filings and uses that data to build ML models that can predict potentially contentious shareholder meetings, allowing shareholders to focus on meetings where their votes may be critical.

On the consumer side, financial institutions such as Affirm, which delivers consumer financing in the form of installment loans through thousands of retailers, leverage ML to power their credit models that make thousands of real-time loan approval decisions each day.

NerdWallet, a personal finance company, provides comparisons of various financial products to millions of customers each month. The company relies heavily on data science and ML to connect customers with personalized financial products. For example, NerdWallet leverages historical data for prospective credit card customers using ML to predict the likelihood of getting approved for a particular card. This allows NerdWallet to recommend credit cards for which a user has a high probability of being approved by the issuing bank.

Reducing risks and fraud

ML engines have also become a key tool for companies to identify fraud and mitigate risk. NuData Security, a Mastercard company, leverages billions of anonymous data points and ML to identify and block account takeover attacks while offering legitimate users a seamless experience. Using behavioral analytics and passive biometrics — such as the user’s browser and IP address, how the person types, how they hold the device and how they move the mouse — NuData can identify whether the right person is behind the device without adding additional friction to the user experience.

Organizations can optimize nearly every aspect of the financial value chain, from front-of-house customer service to back-of-house processes like risk and fraud mitigation. Yet, we’re only scratching the surface. In the years (and even months) to come, we’ll likely see leaders in financial services leverage ML for use cases we haven’t yet imagined.

Alvin Huang is a capital markets specialist for Worldwide Financial Services Business Development at Amazon Web Services, with a focus on data lakes and analytics, and artificial intelligence and machine learning. Prior to joining AWS, he was an executive director at J.P. Morgan Chase & Co, where he managed the North America and Latin America trade surveillance teams and led the development of global trade surveillance.