Can AI solve the engineer shortage?
AI tools are ready to revolutionize software development and help mitigate devastating labor shortages in key areas
The COVID-19 pandemic has put a record 26 million Americans out of work in just a few weeks. At the same time, the virus has put added pressure on industries that were already struggling to fill their ranks. Medical professionals, delivery drivers and many industries deemed essential (like grocery stores and trash/compost removal) are often critically short staffed.
The same remains true for software engineers. While finding enough qualified coders has been challenging for organizations for years, COVID-19 has highlighted an acute shortage in an often-overlooked language—COBOL.
COBOL—short for Common Business Oriented Language—is a programming language that reads like regular English and is often used for business and administrative purposes. Today, COBOL is often referred to as a legacy language since it is no longer used or supported by new systems. However, COBOL still runs many critical pieces of our daily infrastructure, so it remains in use and requires maintenance and upgrades. But as a legacy language, it is getting more and more difficult to find qualified COBOL engineers.
How does COVID-19 affect the need for COBOL engineers?
Many of the systems we rely on every day—some of which have become even more critical during the pandemic—run in whole or in part on COBOL. It’s estimated that 95% of ATM transactions touch a COBOL program, more than 40% of banks still use COBOL as the foundation of their systems and many large government programs—including areas in recent headlines, like unemployment administration and the IRS—rely, at least in part, on systems built on COBOL. Just this month, numerous states experienced significant issues when more than 6 million Americans filed for unemployment benefits in a single week. As systems failed under the stress, emergency requests for qualified COBOL engineers made headlines nationwide.
Couple new and acute issues like this with other ongoing technology skills shortfalls and it’s clear a potential disaster is looming (just what we need in 2020—another disaster). For example, a severe shortage of qualified cyber security engineers is making it easy for bad actors to continue to attack sensitive infrastructures, including those at healthcare institutions. According to a 2019 survey, there are more than 4 million unfilled cybersecurity positions around the world, including more than 500,000 in North America alone. As we all adjust to new normals like working from home and performing more services almost exclusively online, the need for secure systems—and staffs of engineers trained in COBOL and other impacted languages to manage and enhance them—is greater than ever.
This labor problem is twofold. First, of course, we need more engineers — but that takes time. While we’re working to train more software engineers in critical areas, we must also begin to rapidly implement leading-edge tools that can make every engineer more effective. Tools that improve output by orders of magnitude are coming to market, and the implementation of such applications is the output equivalent of doubling (or more) your qualified engineering staff.
Enter AI-enabled software development and management
Artificial intelligence (AI) tools can be extremely effective in improving software development in a number of key areas. According to Tractica, revenue from the application of AI tools worldwide is expected to reach $119 billion by 2025. Specific areas where AI can assist include:
- Automating software troubleshooting. Often, engineers spend days—even weeks—simply looking for areas that need to be addressed, especially when they’re new to a position. AI products can streamline this process and point the engineer to the specific areas of code that require attention.
- Project management. By learning from the actions of engineers, AI can help guide engineers to implement best practices and streamline tasks that help move projects along more quickly. Again, this helps current engineers do far more with the same staff and with higher quality.
- Shortened training curves. Even a seasoned engineer can take weeks to acclimate to programs with thousands of lines of code. An AI assistant can help the coder quickly move from exploration and discovery to making substantive, critical enhancements required to improve the program. As companies can find and hire more qualified engineers, the use of AI-based tools can move them quickly from learning their way around to effectively moving projects forward.
While every business can benefit from improved productivity and efficiency, in any industry that relies on software development, these improvements are critical steps that must be taken to survive. To be clear, COVID-19 didn’t disproportionately impact the engineering workforce, but it certainly turned the spotlight to areas that have required new thinking for some time. When programs are suddenly required to ramp up for requests far beyond any historical measure, things are bound to go wrong. The faster a company can bring existing and new staff members up to speed, the sooner real progress and improvements can be made.
Steve Brothers is the COO of Phase Change Software, a Colorado-based company whose AI-based COBOL Colleague can improve an engineer’s efficiency by at least seven times and currently has beta testing open. For more information, visit https://codecatalyst.ai.