Recently,I've been working on how Generative AI can revolutionize accounting processes, particularly in bank statement processing.

Many companies struggle with attaching invoices to cash flows or automating their financial processes. This often results in manually processing each bank statement line, which is both time-consuming and error-prone.

This processing requires confirming various attributes like VAT, attached business units, payment type, or bundling that need to be split across categories. Labels can be misleading, changing from month to month or involving third-party providers.

It is not something that VLOOKUP could sort, but language models (LM) can.

Using historical data as a reference, LMs predict the categories for bank statements similarly to how a human would process the data: examining each line, understanding its specifics, and finding the corresponding category.

However, unlike humans, LMs can process this information quickly and accurately. By creating a knowledge base from historical data, the LM can query specific labels in a bank statement line, such as third-party company names or specific payment details before making its categorization.

Practically speaking, this AI-driven approach transforms hours of manual work into minutes. From what I've seen, processing 100 bank statement lines manually might take four hours, but with genAI, it can be done in a matter of minutes. The AI performs these tasks without fatigue, and with consistent accuracy.

What is amazing is that this technology is already within reach.

If you’ve been wondering how Generative AI can transform your business processes, data categorization is a clear example. Start exploring these opportunities today and witness the difference it can make.

Hugo Matthaey