Major technology shifts cause disruption and also provide the opportunity for great benefits. Implementing AI for JCL management (AI JCL) will be no different. In researching AI JCL, SEA has found that some AI-related job cuts are beginning to occur. But don’t be alarmed. The situation is more nuanced than it seems and the career outlook for experienced mainframe professionals; especially those with deep institutional knowledge change, may be a lot brighter than the hype makes it seem.
AI is real and it will be used for IBM Z batch processing
Although we might be tempted to think AI enhancement is a bubble that will soon burst or a trend that will soon disappear, reality tells us differently.
Markets are estimating a projected AI investment of $22 trillion over the next five years (figure 1). This is roughly 3.7% of global GDP.
Credible sources are also saying that between 90,000 and 100,000 tech jobs will be eliminated in 2025 alone.
Will I lose my mainframe professional job to AI?
The elephant in the room [and the big mainframe question] is:
Is my role as a mainframe professional at risk?
Current AI-related job cuts are primarily impacting two groups of mainframe professionals:
- New developers doing routine coding
- Entry-level positions focused on repetitive tasks
Experienced professionals who manage mission critical systems are becoming more, not less, valuable. It’s important to understand that while AI can code, it cannot replicate two critical business advantages in JCL batch processing:
- Decades of business rules
- The institutional knowledge embedded in complex batch applications that are orchestrated by JCL
The reality is that these are strategic advantages that mainframe professionals understand, control, and can apply to AI JCL processing. While AI can excel at analyzing data sources related to JCL, it doesn’t know your business rules and the institutional knowledge underlying your JCL batch processing.
That’s the value proposition experienced mainframe professionals bring to the equation. Where AI processing is weak in applying holistic context and multi discipline institutional knowledge; experienced mainframe professionals are strong. AI cannot change that relationship, and there will always be a need for experienced mainframe IBM Z professionals – who are constant learners.
AI strengths and weaknesses for batch processing
To illustrate further, let’s examine what we see as AI strengths in batch processing and what are its limitations, as shown in figure 2.
The month-to-month rate of AI acceleration is impressive, especially from the frontier AI models like ChatGPT and Gemini from Google. And for the IT enterprise, this rate of AI evolution presents a challenge. Any large-scale mainframe architectural initiative, including AI JCL, will face headwinds because the underlying technology is moving faster than traditional project timelines can adapt to.
But these headwinds or the fear that AI will replace your job shouldn’t stop you.
There are many high-value business use cases for AI JCL that are available today. Realizing these benefits will require experienced IBM Z professionals to work cooperatively with AI to create high-value, quality JCL solutions. Solutions that balance AI strengths and weaknesses with the deep expertise and knowledge of IBM mainframe professionals.
What AI JCL will be good at
Looking at the AI strengths listed in figure 2, it’s easy to see how AI can help produce high-value solutions to JCL related issues. AI is good at encoding, indexing, applying, analyzing, and delivering documented organizational knowledge in an easily digestible format. It can assist with error diagnostics and creating corrective actions to resolve issues. It can generate routine code, automate simple tasks, and generate JCL scripts. And as always, AI JCL will excel at greatly accelerating high quality analysis and reporting.
What AI JCL won’t be good at
Now, look at the limitations in figure 2 to see what AI isn’t good at doing. Generally, AI will be weak at assisting with JCL tasks that require deep undocumented organizational knowledge of business needs, rules, and conventions (contextual knowledge) AI JCL will be bad at using unrecorded contextual knowledge that isn’t available for resolving JCL requests and needs, including:
- The downstream impacts of JCL issue resolution on critical applications
- Whether process updates or corrections will ripple and produce negative effects on operations
- How AI JCL error resolutions may affect customer relationships, SLAs, or regulatory compliance
- Whether it is appropriate to make a tough call not to approve a JCL process override during a critical processing window
The value of experienced IBM Z professionals
AI JCL will require human judgement and deep contextual knowledge for IBM Z JCL batch processing. Because of this core need, and the high cost of failure, AI cannot totally replace IBM Z professionals with strong intuitional knowledge for the foreseeable future.
While organizations may be using AI to replace new and entry level, experienced IBM Z professional positions should be secure. Experienced professionals managing mission-critical systems are becoming more valuable, not less. AI can write code, but it can’t replicate decades of business rules and institutional knowledge embedded in complex batch applications orchestrated with JCL. Mainframe professionals’ deep expertise isn’t the problem with AI JCL; It’s the strategic advantage that keeps critical operations running.
More to come
We’re just at the beginning of implementing AI for JCL management. AI-enable IBM Z batch processing is as much journey as destination. In the coming months, we’ll be discussing strategies for fine-tuning your skillsets and developing a practical plan for applying AI to leverage your deep and institutional knowledge to increase organizational value.
Please feel free to contact SEA to discuss your own AI JCL initiatives and to exchange information about incorporating AI into IBM Z business processing.
