Using AI for Large Migrations

I love AI, but until today, I haven't found it saves me time. I usually use it to increase quality of my work, but seldom to speed up my execution. Today I had a task where the opposite happened. My team migrated an API from V1 to V2 and requested all the customers using it to move to the new version. Or course such tasks falls pretty low on their priority. I used AI to write detailed instructions to migrate to the new API, then asked it to run a code search on the import to locate all the potential files to update. I then launched parallel agents to update the code with the goal to limit the number of line updated. It ended up very successful. Most of the pull requests were good without extra edits. Some of them were horribly wrong and needed complete re-dos -- But overall, it was quite brilliant! So, this is now two ways I use AI now. To improve the quality of my work To do easy work in bulk