Financial spreading is the process of transferring information from a borrower's balance sheet into an institution's financial analysis spreadsheet to spot trends and better predict future financial statements. It's a critical part of any underwriting decision — and why
"In a way, it's a giant mapping exercise to ingest all the historical financial data on thousands of companies and automate future updates as they become available," said Katya Chupryna, a director at Citi. "Converting this tremendous trove of information from disparate sources in diverse formats from unstructured to structured creates uniformity and opens up endless opportunities for immediate and future interaction with the resulting datasets."
After Numerated's commercial lending platform receives credit documents from Citi's borrowers through the bank's customer-facing systems, the fintech's artificial intelligence-driven automation tools help compile the statements like tax filings and balance sheets into a singular dashboard accompanied with ratios on liquidity, credit and more.
Underwriters can conduct credit analyses and other reviews of their own within the platform after the data has been validated, before exporting the results to other departments within the bank.
"Through digitization of labor-intensive processes with the help of machine learning, financial institutions can augment manual processes to complete tasks more efficiently and make faster and more accurate decisions with better insights," Chupryna said. The bank also
This initiative is the latest in a wave of tech adoptions by Citi, which has invested in firms like the California-based real-time analytics platform
AI adoption within the financial services space has been gearing up in 2024. According to
One such institution that has built out new tools is the
Other examples include the $3.8 billion-asset
Experts say the growing interest among banks and credit unions in lending technology is largely "driven by the integration of diverse alternative data into their decision-making processes," according to David Donovan, head of financial services at Publicis Sapient, a global digital consultancy firm.
"Lenders now consider a variety of data sources such as payment history, bank account data, educational and employment backgrounds," and more, Donovan said. "These alternative data points complement traditional metrics like credit scores and income, providing a more comprehensive and real-time understanding of an individual's financial habits."
But not all executives are eager to jump into the AI ecosystem. Further findings in the March Arizent report showed that 61% of respondents felt that the technology is evolving too quickly to keep pace with, in addition to a separate 57% saying AI could introduce new ethical concerns or biases into their businesses.
Ian Benton, senior analyst in digital banking at Javelin Strategy & Research, said that while generative AI used internally for assistance with lending is poised for rapid adoption, consumer-facing applications are less abundant.
"Hallucinated data and insights, biases, legal compliance and customer expectations of interacting with a human [all] add up to reasons providers should exercise extreme caution before rolling out customer-facing tools using generative AI," Benton said.
Other hurdles involve data privacy and security restrictions, legacy system inadequacies and the regulatory differences at the state and federal levels over AI usage, said Rajul Sood, global head of banking at the global research and analytics firm Acuity Knowledge Partners.
As Citi's commercial lending experts use the Numerated platform, the bank will continue rolling out the product across its footprint.
"To see a bank like Citi making a global, very significant investment in automating the ingestion of financial statements using AI is a big deal. … Front office automation using AI is going to change it," said David O'Malley, president of Numerated.
