How Generative and Hybrid AI Solve the Technical Debt Crisis for Mainframe Transformation

Anthony DiStauro, an R&D solutions architect at BMC, explains how AI can help to overcome problems that accrue from technical shortcuts

Technical debt has been a persistent challenge across many platforms for decades, but it’s especially critical in the context of mainframes. Mainframe systems power the world’s most essential industries, from banking and insurance to transportation and healthcare. Yet, they also carry decades of accumulated and complex inefficiencies, making them harder to maintain, transform and modernize. But what exactly is technical debt? Why does it exist, and how can we address it holistically? And most importantly, how can generative AI (GenAI) and hybrid AI revolutionize the way we manage and eliminate this costly burden? Let’s explore.

What Is Technical Debt?

Think of technical debt as the price you pay for taking shortcuts in software or operational processes. When developers or operators skip best practices—like avoiding thorough documentation, postponing refactoring or patching instead of solving root causes—they incur a “debt.” Sometimes, business necessities like compliance require quick changes in the short term without the full consideration of long-term implications.

While these shortcuts may speed up delivery or resolution in the short term, they create hidden costs that accumulate over time. Just like financial debt, technical debt accrues “interest,” making future changes slower, riskier and more expensive, causing the cycle to continue in an even more aggressive manner.

In mainframes, technical debt is often magnified. These systems, many of which were built decades ago, have been patched and adapted repeatedly to meet new business demands. Over time, this creates a web of interdependencies, brittle code, dead code and undocumented code, workflows and processes. Additionally, skipping essential software maintenance, delaying upgrades to new versions and neglecting to adopt new features exacerbate the problem. It’s like running a car without a tune-up and only going to the mechanic when it stops running. Add to this the operational complexities of running mission-critical systems, and you have a debt crisis that affects both software development and operations.

Why Do Shortcuts Happen in Mainframe Systems?

The roots of technical debt in mainframes often lie in the intense pressures faced by development and operations teams. These experts often face immense pressure to deliver updates or fixes quickly, especially since these systems power mission-critical functions like banking transactions, airline reservations or insurance claims. Here’s why shortcuts are taken:

1. Time constraints: Mainframes handle business-critical workloads, and downtime or delays are rarely an option. Planned downtime is sometimes possible to satisfy regulators or implement critical updates, but in many cases, organizations demand 24/7 availability. Teams often implement “quick fixes” to resolve issues rapidly, leaving deeper problems unaddressed. The need for continuous operations often limits the ability to schedule maintenance or conduct thorough upgrades.
2. Complexity of legacy systems: Decades of updates, patches and integrations make mainframe environments inherently complex. Developers and operators may opt for workarounds instead of delving into intricate, undocumented code. Additionally, practices that were once considered industry standards have not been refactored over the years, leading to inefficient processes that are now outdated but deeply ingrained.
3. Fear of breaking the system: The tightly coupled nature of mainframe systems means that a small change can have far-reaching consequences. Teams may avoid major changes to minimize risk. Furthermore, the inability to easily test changes in a sandbox or staging environment heightens the fear of introducing errors, leading to hesitation in making meaningful updates.
4. Knowledge and resource gaps: The shrinking pool of experienced mainframe professionals forces smaller teams to prioritize maintenance over optimization. With too few systems programmers (sys progs) and administrators to keep systems running, many organizations struggle to dedicate resources to addressing technical debt or modernizing processes. Instead, their focus remains on “keeping the lights on.”
5. Hardware-driven prioritization: Many mainframe teams spend significant time upgrading hardware every two to three years. This focus on hardware improvements often diverts resources, management attention and budgets away from software improvements. As a result, software maintenance, refactoring and modernization efforts are deprioritized.

The Risks of Technical Debt in Mainframe Development and Operations

The consequences of technical debt extend beyond development inefficiencies. In the operational domain, they create risks that affect the business as a whole:

• Higher maintenance costs: Inefficiencies in both code and operations lead to increased effort and expenses.
• System instability: Fragile systems are more prone to outages, which can disrupt mission-critical services.
• Security vulnerabilities: Outdated processes and code are more susceptible to breaches and compliance issues.
• Slower modernization: Integration with hybrid IT environments becomes a challenge, delaying cloud and AI adoption.
• Team burnout: Developers and operators alike face frustration and fatigue when working in debt-laden environments.

Why Technical Debt Sabotages Future Innovation

Mainframes often represent the backbone of an organization’s IT infrastructure. However, technical debt in these systems can stall innovation:

• Development bottlenecks: Adding new features becomes time-consuming as developers navigate complex, undocumented code. And, as developers address technical debt, this takes up precious resources that could be spent on net new development.
• Operational inefficiencies: Legacy tools and processes slow down incident response times and system scaling, as well as responsiveness to competitive situations or regulations.
• Lost knowledge: Retiring mainframe experts take with them critical tribal knowledge, leaving gaps that further exacerbate the debt problem.

A Blueprint for Addressing Technical Debt With AI

Tackling technical debt isn’t about finding a magic fix; it’s about implementing consistent practices to reduce existing debt and prevent new debt from forming. Here are some foundational strategies:

1. Refactoring: Regularly clean up and optimize code and remove dead code to remove inefficiencies.
2. Automated testing: Implement robust testing to catch issues early and ensure that changes don’t introduce new problems.
3. Continuous documentation: Ensure that all changes to the system are documented in real time to avoid creating knowledge gaps.
4. Modern tooling: Use modern development environments and tools that support automated refactoring, testing and debugging.

While these practices are effective, they can be time-consuming and require significant expertise—two things many organizations are short on today. That’s where GenAI and hybrid AI come in to provide a transformative approach:

1. Rules-based learning for stability and compliance

• Enforces best practices: Codified logic ensures that operational processes and code changes adhere to organizational and industry standards.
• Reduces risk: By automating routine checks and validations, rules-based systems prevent errors that can lead to future debt.
• Streamlines operations: Tasks like compliance auditing and batch job scheduling are automated, reducing manual effort.

2. Machine learning for proactive optimization

• Detects inefficiencies: Machine learning models analyze historical and real-time data to pinpoint patterns of inefficiency in both code and operations.
• Predicts failures: Proactive alerts enable teams to address issues before they escalate into system outages.
• Optimizes resource allocation: By understanding workload patterns, machine learning helps teams allocate computing resources more effectively.

3. GenAI for simplification and knowledge transfer

• Code explanation: GenAI tools simplify complex, unfamiliar code into understandable language, empowering developers to refactor, write new code and refactor existing code with confidence. Many tools also integrate runtime visualizers, providing dynamic insights into code behavior during execution. These visualizations help teams understand dependencies, identify bottlenecks and prioritize optimizations more effectively.
• Automated documentation: GenAI creates up-to-date documentation for code and operational processes, eliminating gaps caused by tribal knowledge loss.
• Incident resolution assistance: AI analyzes logs and provides clearly explained insights with next-step actions to accelerate issue resolution and minimize downtime.

By combining these AI-driven techniques with foundational practices, organizations can comprehensively address both the development and operational aspects of technical debt.

A Unified Approach to Mainframe Development and Operations

The strength of hybrid AI lies in its ability to address technical debt holistically, by combining a variety of AI problem-solving techniques to deliver incredibly powerful solutions to complex technical challenges for both development and operations:

1. For development teams: GenAI demystifies legacy code, while machine learning models identify areas for optimization and refactoring. This accelerates feature delivery and reduces time-to-market.
2. For operations teams: Rules-based learning automates routine tasks, while machine learning, deep learning and GenAI improve incident response times and operational resilience.
3. For the business: Together, these technologies lower maintenance costs, reduce downtime and enable faster modernization efforts, keeping organizations competitive.

---

The Positive Business Impact of Addressing Technical Debt With Hybrid AI

By leveraging hybrid AI, organizations can transform their approach to mainframe management and achieve significant business benefits:

• Cost savings: Automation and proactive optimization reduce the resources required for maintenance and operations. Proactive issue detection minimizes downtime, avoiding lost revenue.
• Faster innovation: Simplified code and automated processes free up teams to focus on developing new solutions, innovations and features.
• Increased agility: Predictive insights and anomaly detection streamline decision-making for modernization and transformation efforts.
• Improved stability: Automated compliance checks enhance system reliability. AI predictive analysis combined with monitoring play a key role in preventing system outages. AI agents can resolve issues without the need of human intervention.
• Enhanced productivity: Developers and operators are empowered with real-time insights and natural language explanations, reducing time spent troubleshooting.
• Future-proof systems: Capturing and preserving tribal knowledge ensures that mainframes remain accessible to future generations of developers and operators.

The Path Forward With Hybrid AI

The technical debt crisis in mainframes isn’t just a development problem—it’s an operational challenge and a business risk. But with the power of hybrid AI, organizations now have the ability to tackle this issue head-on. By combining the strengths of rules-based learning, machine learning and GenAI, hybrid AI offers a comprehensive solution that addresses the root causes of technical debt while enabling modernization and innovation. By addressing inefficiencies, preserving critical knowledge and enabling proactive improvements, hybrid AI empowers organizations to modernize their core systems while maintaining stability and compliance.

With hybrid AI, organizations no longer have to choose between maintaining legacy systems or pursuing innovation. It’s the ultimate approach to help reduce technical debt, future-proof operations and unlock the full potential of the mainframe in a modern IT landscape.

---