Automated Spreading: How AI Revolutionizes Data Analysis in Lending

The Challenges of Financial Spreading in the Credit Granting Process

Financial spreading is an essential part of the credit granting process. It means transferring the data from annual financial statements, or other financial documents of clients into the bank’s financial spreading applications. And this must be done in such a way that the correct information is mapped to the corresponding field in the financial spreading template. Only then do they provide a meaningful picture and can be usefully processed further – not least for a well-founded credit decision.

“The larger the customer base, the more challenging the manual processing and accurate evaluation of financial data becomes,” says Saikat Gupta. In the expert talk, he goes into detail on the core problems that arise from this. Only a few aspects are mentioned here as examples.

As expected, manual data extraction is time-consuming and can take 35 to 120 minutes per dataset. Moreover, it is prone to errors. According to Gupta, eight to ten incorrect entries in 100 data points are not uncommon. Especially in peak seasonal times, it is difficult to maintain consistent data quality when there are high workloads yet processing deadlines must be met. This drives up the costs for later quality assurance.

It is not surprising, then, that interest in automating this process is growing. “Banks desire nothing more than to manage the costly financial spreading more efficiently. And rightly so. Because this is a case for automation with high ROI potential,” says Saikat Gupta.

What is Automated Spreading?

Automated Spreading replaces the manual capturing of financial statement data with AI. The credit analysts upload the e.g., PDF files via a interface or a web application.

Subsequently, AI-trained text recognition algorithms identify the relevant data points and add them to the correct place in the database of the financial spreading software used by the bank. The foundation of specialized software for automated spreading is much more powerful than conventional options for OCR (optical character recognition). It not only makes the text of a paper document digitally searchable but, unlike OCR technology, also captures contextual relationships.

“Automating this is attractive from the banks’ perspective but not trivial from an IT standpoint,” emphasizes Saikat Gupta. “In practice, the credit clientele submits their financial reports in all sorts of document formats, as a scanned PDF, in Excel, Word, sometimes as a low-quality image file, sometimes as a protected file, and so on.” Moreover, the financial accounting standards used may differ, and for banks with an international clientele, this also applies to the languages in which the documents are written.

Software for automated spreading must therefore be able to read in and standardize output from very heterogeneous source material. An indispensable part of this is a so-called template-based approach. It allows the captured data to be displayed in different target templates at the appropriate place. These vary, for example, depending on the legal form of the company, but also over time as internal bank and legal guidelines change. Even then, the assignment should still be correct.

Good solutions prepare the data as precisely and quickly as possible. At the same time, they always present how they calculated the evaluations and which bank guidelines were the basis in a comprehensible, traceable, and transparent manner.

The Benefits of Automated Spreading for Credit Decisions

Evalueserve meets all these requirements for automating financial spreading with a solution called Spreadsmart. It can be seamlessly integrated into ACTICO’s cloud-based Credit Risk Platform. This allows the captured data to be optimally interwoven with the entire digital credit granting process. “Banks that have automated their financial spreading in this way report a time saving of up to 80 percent and data accuracy of up to 99 percent,” Saikat Gupta concludes.

The integration of Spreadsmart with the Credit Risk Platform offers extensive benefits beyond automating data collection. ACTICO’s solution contains fully configurable templates for financial data, which accelerates further processing in the credit granting process. This makes it easy to create reports that are either provided to credit analysis teams as a basis for evaluation – or result in fully automated decisions upon request. Because the Business Rule Engine integrated into the Credit Risk Platform allows banks to also design the actual credit granting based on adjustable business rules as an automatic process.

Because ACTICO and Evalueserve pursue a cloud-based approach, the solution can be easily integrated into the bank’s existing IT landscape and with external data sources via interfaces. In this way, banks realize a modern digital workflow from data entry through creditworthiness analysis to the decision on credit granting. Additionally, the results of financial spreading are seemingly available for the calculation of internal bank rating procedures, which must also be included. This article explains in more detail the opportunities for automating other sub-processes.