The Application Got Faster. Your Best Candidates Left First.

Gorick Ng, a Harvard career advisor and UC Berkeley faculty member, tells Axios that job seekers now assume “every rejection is automated” and that “their video-recorded interview won’t ever even be touched.” That is not paranoia. It is a rational response to a process candidates can no longer verify, and it should worry a VP of TA more than any of the volume numbers making headlines this week.

The Axios story frames the problem as candidate burnout: ghost jobs, AI rejections, and a flood of applications nobody has time to read. That framing is accurate as far as it goes. It stops short of the real cost. The candidates most worth hiring, the ones with other options, are the ones most likely to opt out quietly before a human ever sees their application. No hiring report tracks that number. By definition, it never enters the process.

What 291 applications per hire is actually filtering for

Recruiters now process roughly 291 applications per hire, up from roughly 100 just a few years ago, according to recruiting firm Ashby. That volume forces a response: more keyword filters, more automated screens, more chatbots doing the first read. Every one of those tools is tuned to find matches, not to find strength.

In our work with hiring teams, the applications that clear an automated first screen are usually the ones written to match the posting rather than the ones with the strongest underlying case for the role. Volume forces triage, and triage rewards legibility over substance.

Only 21% of recruiters trust their own filter

Here is the stat in the Axios piece that deserves more weight than it got. Only 21% of recruiters surveyed said they were “very confident” their systems were not wrongly filtering out qualified candidates. That means seventy-nine percent stopped short of that confidence. That is not the same as admitting the screen is unfair, but it is nowhere close to a clean bill of health.

Set that next to Greenhouse’s separate finding that 46% of candidates say their faith in the hiring process has decreased, and the picture stops being about candidate mood. It becomes about the reliability of the instrument doing the screening, reported by the people who operate it.

Walking away is a market signal

The reporting also shows 63% of U.S. applicants have now been interviewed by an AI system, and 38% say they have walked away from a job because it included one. The instinct is to read that 38% as impatience. We read this as the market pricing in risk, not impatience. The candidates who walk away first are the ones who still have a live process elsewhere. They have the least reason to stay and wait it out. Attrition at the top of a funnel does not distribute evenly across candidate quality. It selects.

Set that against the 49% of hiring managers who say AI has improved candidate quality. That number measures who survived the screen. It says nothing about who never applied, or who applied once, hit an AI interview, and left before a recruiter ever knew they existed.

A faster front door, a slower close

Indeed’s vice president of corporate communications, Scott Dobroski, told Axios it now takes almost 25% longer to fill an open role than before the pandemic. That is true even as easy-apply features and automated applications make it easier than ever to submit one. Dobroski describes the result as a “huge volume of noise” employers have to sift through, an experience that leaves candidates thinking they have been ghosted.

The article does not draw a direct line between the two trends, but they are showing up in the same market at the same time. Treat them as one problem, not two: the entry point got wider while the screening behind it got no better at telling strong candidates from weak ones.

The metric nobody is tracking

Ng compares the hiring process to a first date: “How a company shows up in the hiring process can tell you a lot about how this company actually operates.” When employers show little effort in being human, he says, candidates start asking, “If this is what I’m seeing on a first date, should I be worried about our long-term prospects together?”

The candidates asking that question first are not the ones with nothing better to do. They are the ones with somewhere else to be. If you only track where application volume is spiking, you are watching the wrong end of the pipeline. What matters is finding where automated rejection and silence are pushing your strongest candidates out before your own team ever knows they applied.

Original reporting: Axios.