VIDEO TRANSCRIPT

‍

If you're a mainframe modernization engineer, you've probably realized by now that AI can't help with the hard part: verifying that the code actually works in production.

Imogen is a mainframe modernization platform that changes that. It creates tests first based on the behavior of the data flows in the existing system, so you can automate verification as part of the code generation process.

The Imogen pipeline starts by parsing the legacy code base and building a data model of every observable behavior the program has. That model is structural, the records, the jobs, and the flow between them, rather than a written summary of the code.

The existing system's behavior becomes the ground truth, defining the tests that guide the development of the candidate system.

If a verification gate denoted by diamond shapes finds a gap, the work routes back and runs again until it closes.

Code generation also runs through committee models, different from the one that built the candidate.

This is because a single model can produce something plausible, yet wrong.

Disagreement between models flags that case for another pass. Passing tests tell you the code works, but not that it is complete.

Imogen also maps every generated test back to the business rules it extracted from the legacy system. So you can go rule by rule and ask whether each one is covered.

Some legacy rules shouldn't carry forward at all because they work around a constraint that no longer exists.

This view lets you make that judgment call deliberately. Up until now, all of the testing has been done locally. Imogen has also been capturing data from the mainframe. That means for the final test, Imogen can run actual production data flows through the candidate system and compare the output against the same workflow on the mainframe: actual inputs and actual outputs from actual runs.

Failures show in red, and the same packet comes back green as the loop closes them out.

By rewriting the new system based on rigorous tests built and verified from the real production data flows on the existing mainframe, you can be confident that what you have works, freeing you up to make changes quickly, easily, and safely.