CIOs across every industry are looking to modernize their existing systems due to a number of pressures. On the one hand, talent skilled in systems that are obsolete or near-obsolete such as C++, COBOL, PL/SQL, old Java frameworks like Struts, is harder and harder to come by, on the other, operational costs associated with their infrastructure (such as Mainframe MIPS, virtualization platforms, etc.) continue to rise. Moreover, when systems have been customized ad hoc over the time, the situation is even more complex as the staff that originally developed the solution may have left the company or retired.
Younger professionals are not trained on these outdated technologies and have little interest in learning how to use them as newer skills are much more in demand. The most sought-after skills for IT engineers today are cloud engineering on AWS, Azure, or GCP; DevOps and container orchestration; modern backend development with Python, Go, or modern Java; frontend engineering with TypeScript and React; data engineering using Spark, Kafka, and cloud data platforms; AI/ML engineering with LLM integration and MLOps andcybersecurity skills focused on cloud security and application security. All these solutions are far more rewarding for younger workers.
As a result, many businesses have been forced to outsource management of legacy services to offshore companies, losing internal know-how. In extreme situations, even tech vendors are no longer providing services for maintaining outdated systems.
Despite being functionally adequate, these applications have become technologically obsolete and as a result they make it difficult or impossible for companies to leverage the advantages offered by cloud and pose compliance and security issues. Moreover, they don’t enable the provision of new modern digital services that today’s users have come to expect to interact with, such as advanced front-ends integrated with AI tools.
Up to very recently, companies typically had four courses of action available to modernize their systems: purchasing a market product to replace the existing application, lifting and shifting the application to a new environment, converting the code using “language translation” tools, or completely rewriting the code manually. Modernization projects are typically seen as an “IT-for-IT” issue, that is complex to justify, especially as the expertise and man-hours required tend to be hard to come by and expensive. Such lengthy projects are not just hard to finance but tend to embed a high risk of human error due to the manual nature of updating reams and reams of code.
GenAI and Agentic AI, however, now offer a completely new alternative that is cheaper, safer and faster. Thanks to a semi-automated industrialized and GenAI-augmented process, code can in fact be transformed multiple times through an iterative approach. This process improves and refines code by step. GenAI thus not only translates the code from old to new programming languages, but also transforms the application’s structure and architecture, while also producing technical documentation and generating test cases.
GenAI provides a powerful business case: typically, the time saved thanks to the speed at which GenAI can provide updated code migration iterations and its accuracy, can bring savings ranging between 30% to 60% compared with traditional migrations, depending on source and target technologies and frameworks, applications architecture, integrations involved and migration type.
Despite their transformative potential, GenAI and Agentic AI should not be seen as a silver bullet to be used in isolation. Applications are built with a high number of interlaced components, so using an off-the-shelf LLM to translate a piece of code is not enough. A successful project hinges on a number of variables, such as experience in GenAI tools, LLM models, prompts, directives, source and target application technologies, frameworks, architecture patterns and specific guidelines. It also relies on an orchestrated, methodological approach combining this specialist expertise with advanced tooling, and solid governance.
Modernisation projects based on GenAI and Agentic AI achieve the best results when they follow a transparently structured methodology, ideally divided into four key phases:
GenAI and Agentic AI are transforming the experience of application modernisation, making it more approachable and cost-effective. Thanks to the efficiency and accuracy they bring to the process it’s now possible to deliver in a matter of months what previously required years, bringing companies up to date with compliance, security and consumer demand without compromising operational stability.
