Skip to Main Content

Modernizing & Optimizing Mainframe Applications With AI


Experts say there are over 230 billion lines of COBOL code actively running in the world today. The rapid pace of digital transformation is pushing those System Z-reliant organizations to accelerate their application modernization efforts.

One of the biggest hurdles to mainframe application modernization is the talent gap for mainframe-specific skills. Companies must also carefully balance the agility benefits of more frequent code changes with the risks such changes involve. Finally, application modernization and creation efforts must remain affordable.

AI represents a giant opportunity to rethink the entire mainframe modernization process. In fact, a Gartner study concluded that humans and AI working together can reduce coding times by 30%.

IBM’s watsonx Code Assistant for Z is an example of how AI can be applied to mainframe modernization efforts. Leveraging a finely tuned large language model specifically for mainframe code, the solution can help companies both expand their developer talent pool and shorten time-to-market.

The watsonx Code Assistant for Z solution is designed with three primary goals in mind:

  • Accelerated Application Lifecycle: By streamlining processes like code discovery, explanation, and refactoring, the tool speeds up the lifecycle of applications, enabling organizations to evolve faster.
  • AI-Generated High-Quality Code: IBM watsonx leverages AI to generate optimized code, improving the overall quality while cutting down on human errors and time spent on manual coding tasks.
  • Flexibility and Rich Interoperability: The tool provides flexibility to customize your modernization journey based on the application at hand, with interoperability between COBOL and Java ecosystems.

When compared to manual application modernization processes, the AI-assisted watsonx Code Assistant for Z offers substantial benefits, including:

1. Increased Developer Productivity & Experience

Modernizing COBOL applications can be daunting for developers, especially those unfamiliar with the language or mainframe environment. IBM watsonx Code Assistant enhances productivity by providing AI-generated code suggestions, automating refactoring, and generating Java from COBOL. Result? Less manual coding effort, improved developer productivity, and a more streamlined experience.

2. Greater Business Agility

With faster refactoring and automated transformation of legacy code to modern platforms (like Java), businesses can quickly pivot to integrate new features and services in their systems. IBM watsonx helps businesses rapidly adapt to market demands and customer needs with less redevelopment effort.

3. Reduced Risk

Transforming a mainframe application comes with inherent risks including the introduction of errors in legacy systems when deploying new technologies. IBM watsonx mitigates these risks by automating the creation of tests for the newly generated Java code. This helps ensure that the refactored applications perform as expected while minimizing disruption to business-critical services.

4. Expanded Talent Pool

The popularity of Java as compared to COBOL means that once your COBOL applications are converted into Java, you have access to a much larger pool of developers. This reduces the burden of finding or training COBOL specialists and provides an environment more familiar to modern developers.

Generative AI-Assisted Application Lifecycle

The watsonx Code Assistant for Z solution streamlines every stage of the mainframe application lifecycle using cutting-edge generative AI technology. Key capabilities include:

  • Application Discovery to reveal the structure and relationships within your COBOL application.
  • Code Explanation that shows how your COBOL code operates and which areas need to be refactored or modernized.
  • Auto Refactoring of Legacy COBOL to Business Services that align with modern architectures.
  • COBOL to Java Transformation with AI-generated suggestions to ensure important business logic is retained in the modernized code.
  • Auto-Generated Test Cases to validate that the new Java code performs as intended and that the modernized system is reliable.

In short, IBM watsonx Code Assistant for Z brings innovation to mainframe modernization projects. By utilizing AI to automate and assist with key tasks like code transformation and testing, the solution accelerates modernization, enhances developer productivity, and ensures that businesses can stay agile while reducing risks. In doing so, it not only supports today’s business needs but also ensures future readiness by opening up mainframe systems to a broader pool of talent and technology.

Interested? Keep an eye out for our upcoming webinar on the topic of mainframe application modernization with AI or contact us for more information.