By Thierry Bonfante, Chief Product Officer, Unqork
For most enterprises today, driving innovation is like running a marathon with a huge backpack full of heavy stones. For every new application you create, you need to maintain hundreds, even thousands, of legacy applications—some of them going back decades. And as soon as your new applications go live, they too start their drift into becoming yet more legacy applications. This constantly increasing drag of legacy systems slowly strangles your organization’s ability to innovate, increase competitive advantage, or deliver customer value.
Think of all the critical functions that still run on COBOL, a 60-year-old language that still powers $3 trillion in commerce daily, including 95% of all ATM transactions and 80% of in-person bank transactions. Mainframes at 70% of Fortune 500 companies are also built on COBOL, including all 10 of the world’s largest insurers. In fact, some 800 billion lines of COBOL are still in use around the world, as ZDnet reports.
Yet every year, the cost and risk of maintaining these legacy apps grows more acute due to:
- Spiraling operating costs: Maintaining and replacing decades-old mainframes and other end-of-life platforms is itself an expensive (and ongoing) requirement. And the risk of failure increases every year they remain online.
- Complex, undocumented, and aging code: The amount of time and expense required to untangle decades-old code and actually understand how it works is enormous.
- Talent shortages: The number of developers who know COBOL is constantly shrinking. Most are engaged in keeping today’s patched-together COBOL apps running. This cannibalizes your ability to invest in the innovative solutions the business needs.
If this is such a big problem, why has nobody solved it yet?
The tech field is full of failed attempts at migrating COBOL applications. The poorly documented state of these apps, the scarcity of talent able to understand this antiquated language, and the enormous risk and expense of migrations leave companies in a state of paralysis.
One McKinsey study found that, on average, such projects come in 66% over budget, 33% past the scheduled deadline, and 17% short on promised benefits. And the worst part is, even the successful projects simply reset the clock, since the new languages used for the migration are themselves gradually turning into the next legacy stack.
Facing those odds, even businesses with deep pockets have opted for a “wait-and-see” strategy where they keep running mission-critical processes on apps built in antiquated languages like COBOL, Fortran, and .NET apps.
To migrate a COBOL application, you must first understand what the application does and how it does it—often a monumental task. A commitment to documentation hardly existed when these applications were originally created decades ago. Neither did coding best practices. As a result, code governing specific functions is often dispersed throughout the code repository.
To untangle this mess, it can take COBOL experts thousands of hours simply to discover the intent of previous developers. And that is even before they can even begin to refactor the old code for a modern platform.
The GenAI + Unqork breakthrough
However, two transformational technologies—GenAI and codeless development—completely change the COBOL migration equation. GenAI shrinks the timelines, cost, and risk of legacy migrations, and a codeless platform like Unqork is the perfect destination for migrated apps, making legacy tech debt a thing of the past and enabling businesses to ingest new and evolving technologies with zero changes to existing solutions.
Speeding discovery and refactoring with GenAI
Rapidly analyzing legacy code, understanding its intent, and refactoring are exactly the kind of task that GenAI is built to do. With the right GenAI tool, you can automate the identification, extraction, transformation, and even optimization of your legacy application in a matter of minutes.
To be clear, GenAI is not a one-click solution that spits out refactored code that you can then instantly deploy. The process is as much an opportunity for humans to assess the right level of modernization for each application as it is a technology migration. However, GenAI does remove an enormous amount of the tedious, time-consuming, and expensive discovery work that once made migration projects so costly, risky, and overwhelming.
And with a codeless platform, you’ll never have to migrate again
For sure, GenAI is going to play a bigger and bigger role in migrations, which leaves the question—what is the best destination for those migrated apps? If your ideal state is never to migrate again, then the answer is a codeless platform like Unqork.