2025 has emerged as the “efficiency and effectiveness year” for federal IT. Earlier this year, COBOL regained attention when the Department of Government Efficiency (DOGE) alleged inefficiencies in Social Security Administration (SSA) systems, claiming to find 150-year-old benefit recipients. These findings were later attributed to COBOL’s date representation format and known data entry errors already being addressed. The core issue remains the knowledge gap between newer talent and legacy systems. With COBOL no longer taught to recent engineering graduates and approximately 30% of SSA’s CIO team eligible for retirement, DOGE’s targeted cuts to these experienced workers is concerning. These systems in addition to systems run by other government agencies provide essential benefits to millions of Americans, and allowing this institutional knowledge to “walk out the door” without a maintenance and upgrade strategy is unsustainable.
The need for system assessment and modernization across federal agencies is not new, so why haven’t the IT teams at these agencies moved forward? Lack of resources, time investment and the scars from previous failures have all influenced the desire to “not fix what isn’t broken.” But time is quickly running out. So let’s take a look at 3 key challenges that agencies have faced when attempting to modernize and see how lessons learned from federal and state government peers may be able to help.
Modernizing custom-built government applications that have minimal documentation
Federal government agencies operate thousands of applications built and maintained with custom code and technical workarounds. At the SSA alone, 3,600 applications span 10 regional offices, each developing unique system maintenance methods. Many of these systems still operate off of legacy programming languages. For example at the SSA, approximately 60 million lines of COBOL code support core business functions, and the IRS has approximately 160 applications written in COBOL code as well as core applications written with IBM Assembler. With inadequate documentation and knowledgeable talent departing, the risk of system failure grows substantially.
Agencies can address this challenge by leveraging automated tools that provide comprehensive, current documentation without needing to rely on support staff. Solutions like Intellisys use AI/ML technologies to reverse-engineer source code in 20+ programming languages, generating artifacts including assessment reports, system diagrams, and business logic representations that help personnel understand system operations. This simplifies not only the maintenance of legacy systems but also facilitates successful government modernization efforts when an agency is ready. For instance, a critical Department of Defense (DoD) recruitment system had been compromised due to unsupported technologies and legacy components, with no subject matter experts available to help. The application was successfully documented and converted to a modern Java application without run-time dependencies in just 6.5 months using the Intellisys platform.
Today, system maintenance and modernization must be considered continuous exercises with the rapid rate of change of technology. That being the case, keeping a set of up-to-date documentation particularly for custom systems is critical to avoid a reprisal of the problems agencies are facing today. Finding a tool that supports automated documentation generation for both modern languages as well as legacy languages allows agencies to use one solution both now and in the future.
Implementing government modernization within strict procurement and budgeting cycles
Government procurement and budgeting cycles can be notoriously difficult. Not only are there multiple hoops to jump through but agencies must submit budget proposals for multi-year projects that may not include insights needed from the preliminary parts of modernization. If those budgets turn out to be inadequate, then the process of explaining the shortfall and putting together a new budget must start. Previously, modernization was approached as a one-time “big-bang” effort where IT teams attempted to transform systems in one go. However with the failures that those efforts generated, most IT teams are now looking to a more “phased” approach – a modernization strategy that breaks the initiative into steps and prioritizes using detailed information to plan and test modernization approaches prior to scaling. One example of a 5 step phased approach is:
A key benefit of phased modernization is the ability to segment application budgets by distinct stages. Agencies can initially fund just an application assessment project to understand their legacy systems before committing to a full project. A State Department of Transportation successfully employed this approach when modernizing business process mainframe applications written in four different languages including COBOL code and C#.NET code. The department was hoping to adopt some modern platforms as replacements for these applications and needed to understand not only the code and business logic but how the logic then consolidated into business rules or business policies, Working with a services partner, they first contracted with EvolveWare for automated documentation and analysis, using these insights to structure their business rules extraction and budget appropriately. Similarly, the Defense Logistics Agency (DLA) began their ERP modernization by analyzing and rationalizing their application portfolio, generating $23M in savings. Demonstrating such interim success metrics also helps secure funding for subsequent modernization phases.
Managing the transition from mainframe while maintaining continuity of critical services
When it comes to government systems, the costs of failures or downtime can have resounding impacts on the economy, the nation’s defense, citizen healthcare access and more. While modernizing these systems is critical, agencies must be able to maintain their current systems until the new systems are ready to be switched in. System updates may be made to the current system during a government modernization project to address security loopholes or changing regulations and policies amongst other reasons. For longer modernization efforts where larger systems or more extensive transformations are involved, this can lead to large disconnects between the modernization work performed and where the source system is by the time the target system is ready, requiring even more work to update the target system itself.
But automation and agile techniques have changed the game here. A state department of health and human services was looking to extract the business rules from their Medicaid claims system in order to move to a modern platform. However, they were aware that changes would be made to the original system during the modernization effort and therefore needed a solution that could keep track of these changes and allow the modernization team visibility into how the changes would affect their work. Intellisys’ Agile Business Rules Extraction (BRE) solution uniquely provides these abilities. New versions of source code can be run through the platform to identify changes from the original source code and to generate reports to see the impact on modernization work already performed. Using these capabilities, the state was able to document and extract rules from 1.3MM lines of COBOL code in 6 months.
There is no need to keep “kicking the can” down the road
As federal agencies face increasing pressure to modernize their legacy systems, the path forward requires strategic thinking and modern tools. By leveraging automated documentation technologies, adopting a phased approach to modernization, and implementing agile modernization strategies, agencies can overcome the challenges of limited expertise, complex procurement cycles, and service continuity demands. The success stories from various state and federal agencies demonstrate that effective government modernization is possible when the right approach is taken. With millions of Americans depending on these systems, the time to act is now—before institutional knowledge walks out the door and the technical debt becomes insurmountable.
