AI-enabled cloud migration and test automation

The Client

A large central government development team exploring practical AI use across the software delivery lifecycle. TXP supported the team across two priority use cases: accelerating a cloud hosting migration and improving test data generation for an automated test suite. 

The Challenge

The client needed to migrate a mature code base to a new cloud hosting environment while maintaining confidence that functionality had been moved correctly. At the same time, the project required realistic test data across multiple file types and scenarios to support automated testing. Existing processes were manual, time-consuming and prone to rework, particularly where PDFs and Excel files needed to be produced accurately and repeatedly. 

The Solution

TXP used Copilot as a delivery accelerator across the migration and test data activities, with all outputs reviewed and validated by the delivery team. For the migration, TXP assessed two approaches, separated repeatable activities from one-off tasks, and identified where AI could add the most value. 

The selected approach used Copilot to generate supporting documentation that guided code generation, improving code quality while creating maintainable documentation for testing and handover. 

For test data generation, TXP refined the prompting approach, starting with one file type and extending it across further formats. By separating scenario design from file manipulation, the team added scenarios more consistently and reduced rework when file implementations changed. 

The Benefits

• Created up-to-date, maintainable technical documentation to support testing, delivery and future handover. 
• Improved the quality of generated code and reduced the number of bugs raised during migration.
• Saved an estimated 2–3 hours per model during development. 
• Reduced test data generation effort by around 4 days per model.
• Delivered efficiencies across approximately 30 models, excluding additional time saved through reduced defects. 
• Helped enable the legacy hosting platform to be decommissioned, reducing ongoing infrastructure costs.
• Saved around two weeks of development effort and approximately two months of testing effort across the project.
• Improved repeatability, consistency and confidence in the automated testing process.