COBOL — short for common business-oriented language — isn’t going anywhere. Released in 1960 and standardized in 1968, COBOL was developed by the Conference on Data Systems Languages to handle business data processing needs, and it got the job done.
More than 60 years later, COBOL remains the core of financial transactions and the foundation of enterprise mainframes. It supports more than 80% of in-person credit card sales and 95% of ATM transactions and is the foundation for over 40% of online banking systems.
The challenge? Despite its versatility, there’s a coming COBOL crunch. Even as companies look to operate their mainframes for many years to come, COBOL experts are retiring. What’s more, their replacements don’t have the education or experience to manage and maintain COBOL operations.
DIVE DEEPER: CDW can help your financial services break down data silos.
Why Is COBOL Prevalent in Financial Services?
Originally envisioned as a stopgap measure, COBOL quickly established itself as the go-to programming language for large mainframes and operating systems. That was especially true in finance, where many institutions continue to operate on IBM mainframe systems.
COBOL code became the foundation of modern business computing, working behind the scenes to ensure mainframe stability and facilitate large-scale transactions. Its engineers were called COBOL Cowboys — experts capable of navigating the cumbersome and inflexible language to produce desired outputs.
These experts are now older and dropping out of the workforce, and the technologists coming up behind them aren’t trained on COBOL because most of the world has moved on. Newer languages such as Java, C/C++, Python and Go are flexible, functional and in-demand, and they bear little resemblance to COBOL, leaving businesses with a host of new skills but a shrinking window to solve old problems.
RELATED: How IT leaders can prevent SQL injections from entering their code.
How Banks Can Close the COBOL Gap
According to one survey, 91% of IT leaders report that expanding their mainframes is a moderate or critical priority over the next year. But 71% say mainframe teams are understaffed, and 54% say they are underfunded.
With COBOL engineers aging out of the workforce and schools opting to teach newer programming skills over older frameworks, enterprises now face the dual challenge of growing mainframes within an environment of shrinking technical skills.
One solution is to support current code with artificial intelligence. Instead of trying to recruit or retrain new staff on COBOL, businesses could use large language models as coding assistants and invest in building AI experts to help keep COBOL up and running.
Another option is to convert current COBOL code to Java, which could both improve code accessibility and make it easier for staff with newer skills to handle maintenance, updates and troubleshooting. Given the sheer amount of COBOL code contained on enterprise mainframes, LLMs could also play a critical role in this process.
Click the banner to sign up for our newsletter and receive more business IT insights.
Generative AI makes sense in closing the gap. By training machine learning models on large data sets and giving them the freedom to evaluate and evolve, LLMs can both sustain current code and improve over time to support COBOL efficiency.
Given the critical nature of financial services, however, how and where companies run their AI infrastructure plays a key role in success. For example, coding assistance LLMs that are fine-tuned or post-trained on an institution’s existing COBOL code base should be run within on-premises, AI-ready infrastructure to ensure optimal performance while mitigating the risk of data leakage.
This creates a need for AI experience to help navigate these changes without sacrificing security. With expertise in both AI implementation and physical infrastructure management, CDW can help finance and insurance enterprises navigate changing COBOL conditions.
This article is part of BizTech's EquITy blog series.
