Legacy IT systems pose significant challenges for organizations worldwide

Legacy IT systems continue to pose a significant challenge for many organizations worldwide. A common requirement for IT professionals is to assist in modernizing these outdated systems by migrating applications from older platforms such as mainframes to newer ones, either on-premises or in the cloud.

The effort to update and streamline legacy IT infrastructure has been ongoing for decades, with various approaches being implemented over the years. These have included web services, microservices, cloud computing, virtualization, and transitioning to x86-based Windows and Linux servers. The ultimate goal is to create more advanced and flexible systems that can better meet evolving business needs.

The trend of modernizing legacy IT has been a persistent one on CIO agendas for many years. However, despite the significant investment of time and resources, results have often fallen short of expectations. This raises questions about whether AI can finally bring about meaningful change in this area.

Research conducted by Kyndryl among 2,000 senior IT decision-makers suggests that so far, AI has not had a major impact on modernizing legacy systems. Only 10% of respondents reported using agentic AI as part of their modernization efforts, leaving many to wonder whether AI will ultimately live up to its promise or simply become another ineffective solution in the ongoing quest for improvement.

Despite efforts to modernize their IT systems, many companies are struggling to achieve their goals. According to recent research, only about half of organizations that have undertaken modernization efforts claim to have seen improvements in their IT operations.

A significant number of respondents also reported feeling behind schedule and facing cost overruns as a result of their modernization projects. This lack of progress suggests that the challenges associated with legacy systems are more complex than previously thought.

The Kyndryl report highlights the changing motivations behind modernization efforts. Cost reduction and legacy escape, which were once major drivers, no longer hold sway. Instead, leaders are increasingly focused on ensuring that their technology estates, a mix of old and new systems, work together seamlessly.

Agentic AI is emerging as a key tool in this effort. By operating autonomously and providing greater visibility into system dependencies, agents can help organizations navigate the complexities of modernization. However, few companies currently have confidence in their ability to understand these dependencies.

As organizations continue to grapple with IT modernization, some are turning to agentic AI as a potential solution. This emerging technology is being used to automate tasks such as mapping dependencies, generating code, and creating documentation - processes that have long been time-consuming and prone to errors.

Early adopters of agentic AI in IT modernization efforts appear to be gaining an edge over their peers. Organizations not leveraging this technology are significantly more likely to fall behind schedule compared to those who are using it. Despite these promising results, however, the challenges of modernization persist.

Many forward-thinking organizations continue to struggle with the complexities of modernizing their IT systems. In fact, some experts describe IT modernization as a chronic condition that seems impossible to shake off. Even companies with ambitious goals and timelines often find themselves stuck in an endless cycle of updates and upgrades.

Former IBM strategist Andy Thurai comments on this phenomenon, pointing out that many organizations are still running mainframes despite the push for modernization. The pace of progress is glacial, with some projects taking 5-10 years to complete. This slow-moving landscape makes it difficult for CIOs to make significant strides in modernizing their IT infrastructure.

The issue of "technical debt" - the accumulation of outdated and inefficient systems - remains a major obstacle to successful modernization. According to Thurai, organizations are often addicted to buying new software without properly decommissioning legacy systems. This creates a situation where AI is touted as a panacea for the problems it was meant to solve, but ultimately fails to deliver due to unsustainable costs.

The pursuit of using AI to modernize legacy IT systems is a costly endeavor that may not yield the desired results in the long run. According to Thurai, former IBM strategist, organizations that have invested heavily in chatbots and other forms of AI-driven architecture are now facing a new challenge: maintaining these complex systems.

These companies are attempting to leverage agentic frameworks as an orchestration layer to map out hidden technical dependencies and generate missing documentation. This approach is speeding up initiatives that would have otherwise required months of manual labor, but it remains to be seen whether this will ultimately lead to sustainable cost savings.

As Thurai notes, the use of AI in modernizing legacy IT systems may simply be a short-term solution to a long-term problem, with today's architectures becoming tomorrow's legacy environments.