Peter’s note
Two months ago, our AI applicant screener gave a candidate an 86. Good, not striking, and his resume carried no brand-name school or employer, so he could have disappeared into the pile. But… surprisingly… his interview was the best one we had that week. I listened to the recording, we hired him, and a month in, I can tell you we got this one right. The score was not wrong about his resume, but it was never going to bring out this candidate’s drive and applied skills the way an interactive conversation did.
The Short List
The volume
The volume problem today is real. Katie Tanner, an HR consultant in Utah, posted one remote role. 400 applications in twelve hours, 600 by the end of the first day, past 1,200 a few days later, at which point she pulled the listing. Three months on she was still working through them. LinkedIn now takes in 11,000 applications a minute, up more than 45% in a year. In her NYT piece, Kessler calls it an applicant tsunami and I have not heard a better word for it.
AI hiring
Sunil and Saraf describe generative AI “undermining the reliability of traditional hiring signals,” starting with how little effort a polished application now takes with the use of AI tools. That puts much more evaluation weight on the first part of the process that AI cannot write in advance: what a candidate explains in a live interactive and layered exchange, with a dialogue centered on the candidate’s experience relative to the actual requirements of the job.
The applicant side
Two thirds of candidates now use AI to apply. A third of hiring managers say they can spot it, and one in five would reject someone for it. Those numbers cannot all hold for long. Principled rejection of applicants for using the same tools we use to evaluate them cannot last if you indeed want the best candidates. Top applicants are using market available systems where they can and this will become only more prevalent.
Candidate experience
63% of applicants have now been interviewed by AI, up 13 points in six months. The number I keep coming back to is a different one: 70% were never told upfront that AI would evaluate them. Greenhouse’s own verdict is blunt, that most AI in hiring today is “making a bad system worse.” They are right about most of it, and the cheapest part of the fix is telling people what is happening. The purpose is not to avoid human involvement or denigrate candidate interaction. Instead, it is to offer candidates a faster path to express themselves and stand out in an increasing mass of application volume and give hiring companies economical tools to do this.
Tell candidates what is in your first round before they enter it: which step a machine touches, what it scores, and who reads it afterwards. Greenhouse found that 70% never get told. It costs you one paragraph in the invitation email, and it is the cheapest trust you will buy this quarter.
End note
How many candidates does your team actually interview for every ten who apply? Most teams don’t have that number on hand. That ratio is where the real cut happens, not in the resume screen, and it’s worth knowing before it costs you a hire like the great one we almost let go.