For many executives considering the potential of AI to modernize their systems, there is an almost magnetic pull to something new and something to be built from scratch. But is it always wise?
In the article “AI for IT Modernization: Faster, Cheaper, Better,” McKinsey & Company warns readers that failed modernization efforts can cost hundreds of millions of dollars. According to an August 2025 IDC report, this may be one reason why organizations spend up to 80% of their IT budgets maintaining outdated systems. As businesses look to a future built around AI, existing and potential technical debt has gone from an IT concern to a significant threat.
Technical debt is not the only threat that business leaders must consider in an increasingly modern enterprise environment. Organizations also face increasing risks, including:
However, these risks are no reason for companies to retain and invest resources in uncompetitive legacy systems. Instead, company leaders must choose a strategy that maintains working legacy systems while building the modern systems needed for the future.
Current status of legacy systems
Legacy infrastructure constrains operations in every major industry. For example, banking platforms process trillions of dollars of transactions every day on decades-old mainframe systems, and the U.S. financial industry still relies heavily on COBOL, a programming language written before many developers were born. Examples like this create a sense of urgency to modernize, as legacy maintenance prevents investment in value-added development.
The cost of doing nothing comes with several disadvantages, including:
- Organizations cannot launch new digital products.
- Businesses are struggling to meet evolving regulatory requirements.
- Top talent avoids companies running legacy systems.
Strategic benefits of gradual AI modernization
Traditional modernization projects often fail because they try everything at once. Organizations attempting a “lift-and-shift” migration may find themselves with no benefits and end up with inefficient applications running on modern infrastructure.
AI-powered modernization promises even more. Cognizant's research report, “A Two-Year Timeline for AI: The Path to Meeting Legacy Modernization Demands,'' includes participation from 1,000 business and technology leaders from Global 2000 organizations. They found that 66% prioritize improving employee productivity through modernization. A healthcare provider used AI to prevent clerical errors and modernize patient records systems by connecting traditional infrastructure to cloud applications. Staff have more time to spend on patient care and operational costs are reduced.
Practical real-world AI integration
Several organizations are demonstrating how AI can bridge the gap between legacy and modern systems without catastrophic disruption.
- goldman sachs. We introduced generative AI (GenAI) to the entire engineering team as a “developer co-pilot.” AI assistants handle repetitive tasks such as generating boilerplate code, writing documentation, writing test cases, and refactoring legacy codebases. As a result, Goldman Sachs improved efficiency by approximately 20%. Operating in a highly regulated market, the company positions AI within a private environment and incorporates code compliance checks to meet compliance and security standards.
- airbnb. Used large-scale language models (LLM) to accelerate the migration of large-scale tests. The company needed to modernize its testing infrastructure across its entire codebase. AI tools automate translation tasks and reduce engineering burden while maintaining test coverage. This approach allows Airbnb to maintain functionality while updating the platform's underlying architecture.
- U.S. Office of Personnel Management (OPM). We aim to modernize COBOL-based retirement systems using AI. The two-year project, which began in 2025, will use AI to convert code from COBOL to modern programming languages such as JavaScript and Python. AI handles bulk translation while human developers verify and adjust the output. OPM conducted extensive analysis, reviewing millions of lines of legacy code and categorizing it by complexity to focus modernization efforts on the most important parts.
These three examples have the following common elements:
- Integrate AI to augment existing systems rather than replace them.
- Seek incremental transformation rather than wholesale replacement.
- Maintain human oversight and validation of processes.
- Maintain functioning systems and business logic while modernizing your infrastructure.
A strategic framework for modernizing with AI
Organizations looking to modernize their legacy with AI should consider a step-by-step approach that gains momentum over time. This is a very expensive undertaking, so you may want to deal with technical debt first. However, companies need to address the basics first.
Phase 1: Building the foundation
Start by focusing on immediate operational pressures while laying the foundation for larger initiatives. Use AI to extract and document business logic from legacy systems. GenAI can efficiently crawl source code, translate it into natural language, and map it to business specifications. This gives stakeholders a detailed understanding of legacy systems and provides a discrete, high-value integration point where AI-generated code connects legacy systems to modern interfaces.
There is strategic value in confronting what you know about your own systems. Existing business logic represents decades of organizational learning, so don't discard it willy-nilly.
Phase 2: Systematic debt reduction
Organizations can now start tackling technical debt systematically. Use AI to translate legacy code into modern languages at scale and validate each transformation before deployment.
This AI-driven transcoding requires fewer experts in obsolete languages. This reduces your organization's risks associated with a lack of expertise in traditional programming languages. However, companies still need people with the skills to verify translations generated by AI. While companies may not need the skill sets required by traditional systems and languages, it is incorrect to say that a skills gap does not yet exist, as organizations seek candidates who can effectively use these new technologies.
Despite the challenges this skills gap poses, remember that the focus of this step is reducing technical debt. Every time traditional maintenance costs are reduced, resources are freed up for modernization investments. Every time deprecated code is removed, there is one less security vulnerability.
Phase 3: Transform and grow
After modernization teams reduce technical debt and accelerate the adoption of new, modern infrastructure, organizations can pursue more ambitious initiatives that leaders considered too risky in the past.
Deploying AI capabilities across an organization can help employees respond quickly to changing customer requirements, launch new digital services, and help companies tap into adjacent markets that were previously inaccessible due to the limitations of legacy systems.
Throughout all these phases, we treat modernization as a continuous process. We prioritize business outcomes over technical perfection, with a focus on governance and efficiency.
Donald Farmer is a data strategist with over 30 years of experience, including leading product teams at Microsoft and Qlik. He advises clients around the world on data, analytics, AI and innovation strategies, leveraging his expertise across technology giants and startups. He lives in an experimental forest house near Seattle.
