How Conversational AI Could Be the Lifeline for the IT Skills Gap

At the heart of most global enterprises lies an unsung hero: the mainframe – a platform quietly powering massive mission-critical workloads every day. From handling trillions in financial transactions to running key government and retail systems, today’s mainframes remain indispensable even as technology evolves around them.

For years, industry conversation centered on whether cloud computing would replace the mainframe. In practice, that prediction has started to walk itself back – a regression brought on by the fact that mainframes continue to deliver unmatched reliability, security, and transaction throughput – often at a cost and consistency profile that’s difficult (if not, impossible) to replicate elsewhere.

For many organizations, this reality has two implications:

• Core applications still perform best on the mainframe.
• Even newer digital services depend on mainframe-based systems of record.

As a result, the mainframe has, if anything, taken on an even stronger role, anchoring hybrid enterprise architecture.  

The IT Skills Gap

While the platform remains stable, it faces a pressing challenge: a dwindling pool of seasoned mainframe experts.

Decades of enterprise reliance on COBOL and other mainframe-specific skills created deep institutional expertise. Today, much of that expertise lies within individuals nearing retirement. At the same time, fewer early-career engineers are being trained in traditional mainframe environments at the pace needed to fully replace that experience.

For operations leaders, this skills gap in IT creates a longer-term continuity risk:

• What happens during a middle-of-the-night outage when the person who “just knows” how to fix it has already retired?
• What happens when newer team members are hesitant to work in critical systems because they don’t fully understand them?
• And what happens to customer trust when resolving an issue takes hours instead of minutes because only a few people can interpret the system’s signals?

These questions illustrate why this growing skills gap poses a risk to continuity. Mainframes are engineered for resilience, but the operational model around them often isn’t. As experts retire, businesses are increasingly challenged to keep these platforms running smoothly.

Why Traditional Answers Aren’t Enough to Solve the Growing Skills Gap

Historically, organizations have approached this challenge in three ways:

• Invest in training and certifications
• Attempt selective migration
• Rely on a shrinking pool of experts

While training is undeniably essential, it takes time. Especially to get to the level of expertise needed to confidently support these complicated platforms with institutional context. Migration, on the other hand, can introduce its own risks: it’s complex, costly, and often constrained by regulation, performance requirements, or deeply embedded business logic. And, as mentioned, relying on a handful of experts increases operational fragility over time.

None of these strategies alone reduce the immediate cognitive barrier to working confidently in a mainframe environment.

What’s needed is a way to make mainframe intelligence more accessible without changing the mainframe itself. This is where conversational AI changes the equation.  

How Conversational AI Offers a Lifeline

Rather than replacing mainframes or the engineers who specialize in them, AI can act as an intelligent interface layer between them. By enabling teams to ask questions in natural language and receive guided diagnostic insight, conversational tools reduce the friction traditionally associated with mainframe operations.

Instead of relying on specialized commands or piecing together fragmented system outputs, operators can:

• Follow a clear path from the first sign of trouble to the likely source of the issue
• See how different parts of the system are interacting in real time
• Understand where the problem is occurring without digging through layers of technical detail
• Receive guided next steps to accelerate resolution

This doesn’t eliminate the need for deep expertise, but it does make that expertise more accessible and scalable. Junior engineers gain confidence faster, and senior engineers spend less time answering repetitive diagnostic questions and more time solving high-value problems. Beyond faster troubleshooting, this level of contextual insight can also support ongoing performance and cost optimization efforts, helping teams make more informed decisions about workload efficiency, resource utilization, and long-term modernization priorities.

For operations teams under pressure to maintain uptime while managing workforce transitions, that shift is significant.  

The Solution: Rocket^® EVA^™  

Rocket^® EVA^™ embodies this approach by bringing governed, conversational diagnostics directly to core systems. It enables teams to interact with complex environments through plain language while preserving the rigor and stability enterprises expect from their mainframe operations.

EVA lowers the barrier to entry and gives teams breathing room – buying time to upskill, cross-train, and modernize intentionally without putting continuity at risk in the interim. Long-term, it amplifies deep institutional knowledge, giving a broader range of IT professionals the ability to support the mainframe at various levels, building resilience into the growth plan, all the way down to the core.

Longer-Term Implications

“Mainframe modernization” is no longer solely about ripping out core systems that still deliver extraordinary value. The conversation is increasingly focused on preserving the platform's strengths while evolving how people interact with it.  

Learn more about Rocket^® EVA^™ and how conversational AI represents that shift: making trusted system intelligence queryable, contextual, and accessible to the teams responsible for keeping the enterprise running.