Financial institutions are using AI-powered solutions to unlock revenue growth opportunities, minimise operating expenses, and automate manually intensive processes. Many in the financial services industry believe strongly in the potential of AI. A recent survey by NVIDIA of financial services professionals showed 83% of respondents agreeing that AI is important to their company’s future success. The survey, titled ‘State of AI in Financial Services’, also showed a substantial financial impact of AI for enterprises with 34% of those who replied agreeing that AI will increase their company’s annual revenue by at least 20%.

The approach to using AI differed based on the type of financial firm. Among fintechs and investment firms, the most cited AI applications were algorithmic trading, fraud detection, and portfolio optimisation. This reflects a primary focus on protecting and maximising client returns. In contrast, banks and other financial institutions noted fraud detection, recommender systems, and sales and marketing optimisation as their top AI use cases. Consumer banks not only focus on fraud detection and prevention, but also build AI-enabled applications for customer acquisition and retention along with cross-selling and up-selling personalised products and services.

From capital markets to consumer finance to fintechs, AI is powering the future of finance. Traders are using AI and high-performance computing (HPC) to accelerate algorithmic trading and backtesting, while meeting industry regulations through explainable models. Fintechs and traditional banks are transforming the delivery of financial services across services and products—such as banking, lending, insurance, and payments—with AI-enabled solutions. And AI is improving productivity for financial institutions through virtual agents in call centres and automated analysis of lengthy financial documents.

In this article, we’ll highlight the role of AI across capital markets, retail banks, and fintechs with real-world examples of AI in action.

Building the AI-powered bank

AI is enabling incumbent financial institutions to deliver smarter and more secure services to their clients and customers. Take Royal Bank of Canada (RBC), for instance. RBC built a private AI cloud for banking to run thousands of simulations, train AI models, and analyse millions of data points in a fraction of the time than it could before. The private AI cloud has helped reduce client calls and resulted in faster delivery of new applications for RBC clients. As a result, RBC expects to transform the customer banking experience with a new generation of AI-enabled smart applications.

Firms are also using AI solutions to create robust fraud detection and prevention systems, accelerate risk calculations and fraud detection. BNY Mellon, one of the world's largest cross-border payments service providers that processes more than $1 trillion daily, built a collaborative fraud detection framework that runs Inpher’s secure multi-party computation — which safeguards third party data. The bank’s ML and AI models were trained on over 100 million data samples, and improved fraud prediction accuracy by 20%, while preserving the privacy and residency of the input training data.

Accelerated computing for traders

Market data volumes have surged with the emergence of new instruments, data types, and venues. To stay competitive, financial institutions are bringing the power of AI and HPC to adapt to real-time market conditions and shortened trading windows. Successful trade execution is often measured in nanoseconds, and faster computing results in smarter trade strategies and increased opportunities for profit.

Building end-to-end trading infrastructure that combines enterprise AI with high-speed networking is key to provide the lowest latency and highest bandwidth trading. Trading firms are scaling out with Ethernet switches, adapters, and messaging accelerators to accelerate every point in the trading cycle. Discretionary and systematic traders can be augmented with teams of AI assistants to squeeze more intelligence out of target windows to optimise trading.

Protecting payments with AI

Payments power the global economy, whether transferring money to family and friends, paying bills or buying products online, or using your phone to check out in-store. Financial firms are using AI to improve security and transparency in systems for payments fraud detection and prevention, as well as for identity verification to meet regulatory requirements associated with Anti-Money Laundering (AML), and Know-Your-Customer (KYC).

American Express, for example, uses fraud algorithms to monitor every transaction on their platform in real-time for more than $1.2 trillion spent annually. The financial giant deployed deep-learning-based models to detect fraud and generate decisions in milliseconds.

Creating accurate insurance policies

AI enabled applications are significantly impacting the insurance industry as well as insurers move beyond traditional claims management and embrace digital workflows that employ a fully analytics-driven approach. This includes using AI to automate claims processing, to identify fraudulent claims and to create new digital services to increase customer satisfaction.

For instance, Cape Analytics is a computer vision startup that transforms geospatial data into actionable insights for insurers to write better policies and provide suggestions for homeowners to protect their property against wildfire damage. The startup uses AI to produce detailed data on the vegetation density, roof material and proximity to surrounding structures along with a calculated risk that homeowners can use to take preventative action. Cape Analytics trains its models on servers and uses them for live inferencing, with geospatial data converted into actionable structured data in seconds.

Fintechs use AI for disruptive innovation

Fintechs are creating more intuitive and personalised interactions between customers and their finances using recommendation engines, conversational AI, and deep learning fraud detection models.

NerdWallet, a fintech focused on personal finance, uses machine learning in its recommendation engine to match its customers with the best-fit financial products, such as mortgages and insurance. The fintech’s models learn how profile features including credit scores, outstanding balances, and credit utilisation are getting members approved or declined. As their models become more familiar with underwriting procedures, they improve their ability to match NerdWallet’s members with suitable products.

Square, a financial services and digital payments fintech, uses conversational AI to power its virtual assistant that understands and provides help for 75% of customer’s questions, and reduces appointment no-shows from potential customers with sales teams by 10%. Their team uses a mix of small, medium and large NLP models, and is working towards a general purpose NLP model in the long-term. As Square Assistant expands from dozens to thousands of tasks, its neural network models expand to handle more requests from small business customers.

Whether in accelerated trading, automated call centres, real-time fraud prevention, or other financial services, AI is helping financial institutions drive the future of finance for their customers and clients. Ultimately, financial institutions will AI-enable hundreds, if not thousands, of applications. Those banks that invest in enterprise AI transformation stand to gain market share, improve customer satisfaction and improve their financial performance at the expense of those that fail to innovate in AI.

Learn more about AI in finance by downloading a full copy of NVIDIA’s survey report ‘State of AI in Financial Services’ here.