Legacy Code Modernization

AI helps developers understand legacy code, generate modern replacements, and validate the changes before release.

Business impact

• Manual coding hours — Morgan Stanley reported saving in excess of 1 million manual coding hours.
• Legacy code modernized — Morgan Stanley reported more than 17 million lines of legacy code modernized.
• Developer output — CGI reported 30-40% more output after three months of ramp-up.

Data requirements

• Legacy source code (text) — The tools work on older application code that needs modernization.
• Documentation (text) — One account says developers reverse-engineered documentation for legacy COBOL and Fortran code.

AI methods and techniques

• Code explanation — The AI explains legacy code so developers can understand its behavior and business logic.
• Code generation — It produces replacement code in modern languages or helps draft modern implementations.
• Reverse engineering — It infers structure and logic from older codebases to support modernization.
• Testing support — In one account, the output is tested and compared in a sandbox before production.
• Design generation — One account says the system produces designs that developers can use to migrate code into updated architectures.

Industries

Real-world evidence

5 documented case studies on record.

Companies using this: Caterpillar, GM, JPMorgan Chase, Morgan Stanley, CGI.