Upfront bias: I've built a tool that tailors r sum s to postings and it launches here Tuesday. Which puts me on one side of a question I keep going back and forth on.
Against: if every application is AI-optimised, keyword matching becomes noise, recruiters get more volume, and the signal from "this person clearly wrote this for us" disappears.
ATS was already a keyword filter before AI arrived. People who knew that gamed it; people who didn't got screened out over formatting. Tooling just levels that.
Where I landed while building: it should only use what's genuinely in your experience, and it should tell you which lines it couldn't evidence. Mine flags them and makes you confirm before sending. I'm not certain that's enough.
