Modernisation in Action: Turning Risk into Opportunity with AI

Modernisation isn’t just a tech problem. Most challenges (and solutions) come down to people and processes. That’s one of the insights from our recent fireside chat with Patrick Machado, CTO at Critical Software, who shared lessons from decades of helping organisations modernise their most critical systems, and Leopoldo Andres, Principal zStack Technical Sales Manager at IBM. 

For over 15 years, Critical Software and IBM have partnered to deliver modernisation across industries — from banking and telecoms to insurance and government. Together, they’ve developed solutions that manage vast networks, prevent system failures, and harness AI to streamline operations. 

Yet, even with cutting-edge tools, the greatest risk in modernisation isn’t technical: it’s hesitation. Too often, projects stall before they start because the fear of disruption outweighs the drive for progress. As Patrick notes, a strong business case and executive sponsorship are essential foundations for success. 

When modernisation projects do launch, technical hurdles abound. Legacy systems can be decades old, with lost documentation or even missing source code. The result? High technical debt, complex dependencies, and underestimated risks. Critical Software’s approach focuses on de-risking from the start: understanding the business context, mapping dependencies, gradually modernising systems, and applying rigorous testing at every stage.  

AI is becoming a key enabler in this process, particularly in understanding legacy code. IBM watsonx, a next-generation AI and data platform, supports modernisation by using generative AI to interpret, refactor and optimise legacy codebases such as COBOL, reducing manual effort and uncovering hidden dependencies. Similarly, CoBot, Critical Software’s own AI accelerant, complements this by automating code analysis, documentation, and transformation tasks — helping teams modernise faster while maintaining full control over quality and compliance. 

But as Patrick emphasises, AI isn’t a silver bullet. Generative models must be validated and guided by human expertise to ensure accuracy and maintain trust. 

Ultimately, successful modernisation is a balance. Companies must weigh the risks of inaction (slow time-to-market, skill gaps, outdated systems) against fear of disruption. Step-by-step modernisation, solid testing, and the right partners make the difference.  

“The riskiest project is the one that never starts. Build a solid business case, manage risk intelligently, and choose partners who understand critical systems.” — Patrick Machado, CTO at Critical Software.  

Modernisation is challenging. But done right, it’s transformative.  

Dive deeper and read Modernisation Reimagined with AI, our pocket guide to safer, smarter modernisation, developed in partnership with IBM.