Altisource
Altisource needs to re-engineer its business applications for more than 10 years to support new real estate and mortgage operations, but the legacy of Java has a large code volume and its context spread across Jira, Bitbucket and design documents, and the traditional re-engineering speed is difficult to meet the business window.
The company incorporated the generative AI into the enterprise-level modernization method, allowing Amazon Q Developmenter to re-configure, interpret and change recommendations in combination with the entire code library and Jira, Bitbucket and the design context; the team continued to use Scrum, testing, secure scanning and business demonstration to control delivery and validate effects with story points, gaps and application delivery.
AI is responsible for understanding the legacy context, generating and re-engineering codes; engineers review design and codes, running tests and safety checks, and product and business owners take receipt and inspection functions and decide to publish them.
- Accessing tools to complete development context instead of individual document completion
- Selecting legacy modernization as a clear operational objective
- Maintenance of Scrum, Review, testing and safe access
- Authentication with real application instead of sample code
- Simultaneous measurement of changes in speed, output and gaps
- The code Agent value comes from the context of the warehouse and the project
- Productivity indicators must be listed alongside quality and safety indicators
- To demonstrate the state of modernization of applications
- AI generation code cannot bypass existing engineering doors.
The AWS case revealed that team modernization exceeded 350,000 lines of Java code, with a 25 per cent increase in development productivity, four new applications delivered in four months and a 54 per cent reduction in code holes; the three-month average of Scrum story points went from about 160 to 200.
- 350,000 lines, 25 per cent, 4 applications, 54 per cent were from AWS client cases.
- The story point is not suitable for direct cross-team comparisons and can serve only as a clue to the changes that take place before and after the team.