About the evening
Legacy migration has always had two hard problems. First: understanding what you're actually replacing. Decades of undocumented behavior, embedded logic, and edge cases exist only in production, and most teams skip properly analyzing them because the analysis is too expensive. Second: proving you got it right. Not just "does the new system work?" but "does it work the same way?" Without the legacy system's actual outputs as a reference point, you're not verifying equivalence, you're verifying your assumptions about it.
AI changes the economics of both. Things like multi-source triangulation, AI-generated behavioral specs, and characterization testing at scale are turning what used to take months into days, making the verification problem actually solvable. On June 17th, Equal Experts is hosting a session in New York to walk through the techniques, the trade-offs, and what it takes to make this work in practice.

