Key outcomes
By accelerating mainframe modernization with GenAI, this major retailer has:
Turned legacy code into clear business logic 80% faster
Created a strong foundation for a more efficient supply chain
Proved the value of GenAI for rapid enterprise modernization
A strategic mission to modernize supply chain operations
Like any multinational retail business, this leading grocery and general merchandise retailer must manage a vast, complex supply chain and distribution network to bring its customers affordable, high-quality products.
To enhance its customer offering while optimizing costs, the company embarked on a strategic initiative to streamline supply flow and enable efficient last-mile delivery. However, to achieve this ambition, it needed to rebuild core legacy systems that were rooted in aging technology.
The legacy code challenge: Increased risk and barriers to modernization
Two critical systems essential to the company’s supply chain were based on decades-old mainframe technology.
The supplier order management system, which processes millions of orders daily, was built on an obscure mainframe technology that only a few experts could understand and maintain. Similarly, the company’s warehouse management system (WMS) was built on COBOL with limited documentation, making it extremely challenging to maintain and modernize.
With a scarcity of mainframe code experts to maintain existing systems, the company faced significant risk and major roadblocks to its modernization ambitions. They partnered with Thoughtworks to help understand and extract the core business logic of the legacy code — leveraging the power of AI/works™ to accelerate its supply chain and distribution modernization program.
Solution: Accelerating mainframe modernization with AI
Thoughtworks began by collaborating with the company on a proof of concept (PoC) to test whether generative AI could translate the mainframe code in the supplier order management system into clear business logic.
Deploying our AI/works™ platform, we rapidly interpreted the legacy system rules into fresh, modern specifications — using Code to Spec, the reverse engineering capability within AI/works™, to deconstruct the legacy codebase and surface the underlying business logic as clear, actionable functional specifications.
To further aid comprehension, we created knowledge graphs to visualize code dependencies and generate transaction summaries, giving both developers and product teams a complete, navigable picture of the legacy landscape.
The PoC quickly demonstrated success, achieving its first milestone in just three weeks, and importantly, building leadership confidence that AI-powered modernization was achievable.
To build on this success, the company continued to collaborate with Thoughtworks to tackle the larger WMS modernization challenge of extracting and clarifying the business logic embedded in the COBOL code.
Within three months, the team delivered an MVP of an AI-powered modernization accelerator that could extract business logic and present it as natural language explanations and flowcharts. Thoughtworks also developed a persona-driven chatbot to provide accurate, context-aware answers to questions, translating the complex codebase into clear business intelligence.
legacy code analysis, accelerating modernization efforts.
during the project’s first milestone.
in business logic explanations for COBOL systems.
Outcomes: A foundation for supply chain efficiency and future innovation
The initial PoC with supplier order management system reduced the time required to understand the codebase rules by 80% compared to manual effort, resulting in a 90% cost savings during the project’s first milestone.
The WMS MVP also delivered immediate value, with an 80-85% reduction in the time required to understand legacy COBOL code and 100% factual accuracy in the business logic explanations.
This AI-powered modernization project has also created significant strategic value for the company. By laying a foundation for a more efficient supply chain, shoppers can look forward to improved product availability and delivery. What’s more, it has provided a solid foundation for further AI initiatives and future-proofing.
Next, the company will continue partnering with Thoughtworks to expand and accelerate its AI initiatives in other areas of the business, such as warehouse operations, cybersecurity and anomaly detection across finance and HR functions, building composable capabilities that will allow them to stay modern even as regulations, models and business priorities shift.
