Banks to reap rewards of GenAI

Dimitrios Papanastasiou, Head of Risk and Finance Solutions, and Yasman Moghaddams, a Risk and Finance Analytics Solutions Specialist at Moody’s Analytics

Dimitrios Papanastasiou, Head of Risk and Finance Solutions, and Yasman Moghaddams, a Risk and Finance Analytics Solutions Specialist at Moody’s Analytics

Banks are already starting to see the great economic benefits of generative artificial intelligence (GenAI), as it increases efficiency, accelerates processes, and improves services in areas such as marketing, fraud detection, and credit scoring.

The technology will boost banks’ productivity and could increase the sector’s annual revenue by $200-$340 billion and boost operating profits by 9% to 15%, according to a McKinsey study. In pilot tests, users of Moody’s Research Assistant GenAI tool, which generates insights from Moody’s credit research and data, were able to cut their time spent collecting data by up to 80%, and the time needed for analysis by as much as 50%. The implementation and application of GenAI is therefore not a question of “if”, but “when” and “how”.

What are the applications?

Banking, with its complexities and heavy reliance on data, is especially suited to GenAI, allowing us to re-imagine the way we view data and resource-intensive processes. Imagine faster and more targeted sales efforts resulting in higher client satisfaction. This means not only more revenue for banks through personalized customer engagements, but also reduced errors, more efficient use of resources, identification of new opportunities, and better data-driven decision-making and risk management.

The technology can enable and enhance a wide range of banking tasks. GenAI can augment traditional credit risk assessment with rich information and analytics based on a much broader dataset, ranging from behavioral, compliance, and know-your-customer (KYC)-related information to news reports, outlooks, sustainability/climate assessments and even cyber risk and supply chain indicators for a truly integrated view of risk.

"The technology will boost banks’ productivity and couldincrease the sector’s annual revenue by $200-$340 billion andboost operating profits by 9% to 15%, according to a McKinsey study."

Al-based processing can automate spreading – many banks are already benefiting from applications extracting key information from a borrower's financial statements – with tools such as Moody's QUIQspreadTM. Models can also identify outliers and abnormal patterns in financial statements and can spot possible indications of misstatements. Customer information can also be sourced from accounting software, tax returns, credit bureau, or third-party master and financial entity databases, such as Moody's Orbis Global Entity data. When properly integrated, GenAI can orchestrate the process and synthesize the information.

Faster, better results

GenAI is an excellent tool to improve research for faster processing of more data from a wider range of sources to support decision-making. Banks can gain a more holistic view of a company, a group of companies, or a sector by considering not only the credit quality, but also different types of risks that may become key drivers of financial performance and credit quality in the future.

Analysts can intuitively interact with chat-like user interfaces such as Moody’s Research Assistant, to query internally or externally available content, combine insights across market segments, and put together comprehensive reports or concise summaries to present the research output.

For most banks, generating credit memos and notes is a manual, time-consuming process. GenAI can pre-generate the bulk of the memo, incorporating quantitative and qualitative information, and tailoring output to match standards.

The technology can also generate early warnings, based on multiple sources of information, combining financial performance, behavioral aspects, macro and industry outlooks, and supplementary information from news sources. GenAl models trained on historical performance data can typically uncover patterns and provide better signals to identify problematic areas in a portfolio.

Once an at-risk borrower is identified, the model can suggest how to prevent further deterioration of a loan, for example, by adjusting credit limits, restructuring the debt, consolidating it, or connecting the customer with financial advisors. The model could also generate customized communication to borrowers via email, text, or phone with timely recommendations or warnings.

Looking to the future

The industry is quickly recognizing the advantages of GenAI and its applications. Leading banks have already started integrating GenAI into their organizations’ DNA in the applications mentioned - and there is much more to come. The potential is limitless. Still, for a highly regulated industry such as banking, there must be a balance between speed of deployment and the need for proper testing and governance to ensure robust, accurate, and unbiased input and results. Hence regulation is expected to drive the way many organizations leverage GenAI and ensure transparency.

More fine-tuning is needed to develop and improve the quality of GenAI models, and specialized skills in both prompt engineering and Al model training are in high demand. These skills differ from the risk and statistical modelling skills typically found within banks.

Data centers may require updates. Emerging chipset architectures, hardware enhancements, and optimized algorithms will also be significant contributors. Cloud infrastructure will become vital for deploying GenAI efficiently and sustainably, considering expenses and environmental impact.

Financial

Risk Management

Supply Chain

Artificial Intelligence

Fraud