Can AI Cure the Curious Case of COBOL in the Finance Industry?

The banking industry is making concerted efforts to rejuvenate and alter entrenched perceptions of there being too many slow-paced organizations. However, despite many positive signals that banks are adopting disruptive technologies to get them ahead of the curve, many are still hampered by having to rely on a programming language that is over 50 years old. 

COBOL was developed in the late 1950’s with an emphasis on commercial, finance, and administrative applications. It is one of the oldest programming languages in computer history still in use – despite the development of more modern, flexible languages. However, a solution could now be in sight to deal with this legacy courtesy of AI and Large Language Models.  

The significance of COBOL in the finance industry cannot be overemphasized. More than 43% of international banking systems still rely on it, and 92% of IT executives view it as a strategic asset. More than 38,000 businesses across a variety of industries, according to Enlyft, are still using COBOL. Not surprisingly, it is difficult to replace. 

A large percentage of the daily transactions conducted by major companies such as JPMorgan Chase, American Express, Fiserv, Bank of America, and Visa rely significantly on COBOL. Additionally, some estimate that 80% of these financial giants' daily transactions and up to 95% of ATM operations are still powered by COBOL.  

The Urgent Necessity for Modernization 

Although these companies have been successfully maintaining their COBOL systems for years, efforts to update this software are hampered by trouble locating developers fluent in these archaic languages.  

We know that many British banks, for instance, are battling with several IT issues originating from outdated technology. These include challenges like blocked account access and payment processing glitches.  

This emphasizes how urgent it is to switch from COBOL and other old systems to more modern ones to maintain the financial sector's smooth and safe functioning in a rapidly evolving technological environment. Reliance on COBOL also makes it extremely difficult to integrate the new technologies and apps that are now critical to controlling costs, beating new competitors, and meeting higher customer expectations.   

Integrating Modern Software 

Banks are therefore working with technology partners to create a product that will help them accomplish three key goals: validating, converting, and restructuring COBOL code.  

Recent developments show that Large Language Models can be a perfect solution for this. With the ability to comprehend and interpret code produced in a variety of programming languages, these LLM’s can act as a code helper. They are being used to translate code written in more contemporary languages into older ones, like COBOL. 

Competitive Edge 

Developers then play a crucial role in ensuring the accuracy and functionality of the newly generated code after it has been automatically transformed by the LLM. The advantages of AI are combined with human knowledge to ensure the success of this procedure. 

It is anticipated that the application of AI models in this process will greatly lower technology costs, strengthen security, and speed up development. It has been observed that the usage of code assistants may help reduce around 10% of the expenses that are generally attributed to technological costs in a typical bank. 

The goal of introducing code assistants is to advance a bank's client services as well as its internal IT infrastructure.  

So, it looks like modern software might not only help banks become more efficient, secure, and customer-focused, but also offer a solution to the COBOL headache. All this can then combine to give them a new competitive edge in the banking sector.