It’s failure in a fancy suit
I keep running into the same requirement when organisations replace or upgrade systems: “Everything should work as before.” It is often the only thing people agree on. It sounds simple, but it is not.
I have seen projects where “as is” would require hundreds of hours of development and testing, not because the system was complex, but because the existing behaviour was never clearly understood. Worse, “as before” often turns out to be wrong, full of workarounds, hidden calculations, and manual steps no one trusts but everyone relies on.
In one case, preserving behaviour would have meant rebuilding broken logic. By asking how it actually should work, we uncovered fifteen critical bugs and removed an entire manual calculation. The system became simpler, safer, and easier to test.
“It should work as before” makes testing almost impossible. What are we verifying, every workaround, every undocumented exception? The phrase hides uncertainty and pushes the cost downstream.
I don’t hear stubbornness. I hear fear, and people avoiding responsibility for improving how work actually gets done. If we want safer systems and better outcomes, we have to stop hiding behind “as is” and decide what we are no longer willing to accept.