The role of generative AI in mainframe modernization

By: Kent Ramchand

Mainframes have long been the backbone of enterprise computing, delivering unmatched reliability, security and scalability for mission-critical workloads. Even with the rise of new technologies, these systems continue to power essential industries, including healthcare, finance and manufacturing.

However, as businesses leverage the scale and flexibility of cloud environments, the need for mainframe modernization becomes clear. The rush to explore and adopt generative AI has only accelerated the case for modernization.

Kyndryl’s 2024 State of Mainframe Modernization research found that 86% of respondents are deploying or planning to deploy generative AI tools and solutions to their mainframe environments. Less discussed, however, is the potential for generative AI to help unblock challenges that have, to date, held back mainframe modernization ambitions. While much attention has been given to the promise of generative AI at the application level, it also has potential to play an important role in the modernization process itself. 

Streamlining the modernization journey

With a legacy dating back more than 60 years, mainframes can offer organizations decades worth of data.

But before they begin to unlock the insights in their mainframe data, organizations have to map a path to modernization. For many, this is a challenging undertaking, requiring them to not only create access to the data, but also understand the applications that use it and the business logic that drives them, before developing plans to transform the data and applications to work with modern architecture.

Further complicating matters, mainframe skills and knowledge can be difficult to find today. For many organizations, workers with mainframe expertise and skills in languages like COBOL have retired and taken their knowledge with them. Many businesses rely on external expertise to maintain mainframe operations. (The State of Mainframe Modernization research found 77% of organizations have skills gaps that have only been solved with the help of partners.)

But modernization requires more than business-as-usual operations and maintenance. Tasks such as unpacking and understanding legacy applications are often made more complicated by documentation being incomplete or lost entirely, for example.

For many organizations, even before they get to AI-powered applications, developing the requirements for a modernization project can be challenging, time-consuming and expensive.

Enter generative AI.

AI-assisted modernization

Generative AI can be used to inspect applications and produce — or reproduce — documentation. This function is particularly useful in the case of old applications with documentation that is incomplete or unavailable.

If the generative AI model has been trained on a diverse range of applications and projects, missing information can be inferred from other similar projects. Partners who have expertise and data built up across multiple customers and projects can be valuable here, bringing in sources of information that go far beyond an organization’s own set of systems and applications.

Similarly, business and data logic can be mapped with assistance from generative AI, helping understand both the workings of older applications and the opportunities that could be leveraged by incorporating that data into modern platforms.

Decision support systems that help determine the most appropriate applications to modernize can also leverage generative AI, helping avoid the trap of running projects that consume resources and add to cloud workloads and costs but deliver little discernible business benefit.

While it is an emerging use case, early work by Kyndryl has shown that leveraging generative AI to produce requirements can dramatically cut the timeframes required for mainframe modernization, and significantly reduce the time an organization’s own people need to invest in the requirements-gathering process.

This streamlining helps both lower costs and bring forward the value that the projects are ultimately designed to deliver.

Staying secure with generative AI’s help

Many mainframe applications were conceived and coded before the internet era, which makes security a key consideration for any modernization project. The 2024 State of Mainframe Modernization research found 92% of respondents said regulatory compliance influences their modernization decision-making.

Here again, generative AI can play a role, enabling data encryption and significantly enhancing capacity for security testing and assessment, as well as automating tasks like code modernization and inspection, test case development and automated configuration.

By analyzing large datasets in real time, generative AI can also play a vital role in fraud prevention, providing insights that help organizations meet industry standards and protect against financial losses.

There are many more areas in which generative AI can help businesses adapt to new needs and keep core mainframe, whether by automating routine tasks, reducing manual intervention, cutting processing times or lowering costs. 

Generative AI also can free teams from routine, manual tasks like spending weeks on translating and testing code to concentrate on more strategic imperatives, like aligning applications and performance to broader IT strategies and business goals.

Traditional strengths, modern opportunities

Decades after they were first introduced, mainframes continue to power mission-critical business functions and deliver security, reliability and performance. But many organizations now also see the advantages of leveraging those strengths in concert with modern applications and architectures.

Kyndryl’s 2024 State of Mainframe Modernization research shows the business payoffs to this approach can be substantial. Businesses that modernized workloads on the mainframe reported one-year ROI of 114%, while those that integrated mainframe with other platforms, such as cloud, reported even bigger gains, with ROI of 145%.

As organizations again reimagine the role of mainframes in the generative AI era, AI not only is shaping up as a game changer for data and applications, but as an increasingly important part of the modernization journey itself.

Kent Ramchand is Director of Customer Technology Advisory for Kyndryl Australia.