Why Faster AI Screening Hasn't Made Hiring Faster

Suppose you run a recruiting team that adopted AI screening last year, and it worked. Shortlists that used to take twelve days now show up in three. You'd expect time-to-offer to drop by about the same amount. At a lot of companies it barely moves, and the reason is worth understanding, because it changes what you should try to improve next.

The shortlist has to go somewhere. It goes to a hiring manager, whose week looks nothing like the recruiter's. Maybe three interview slots are free. There's an all-hands on Wednesday, board prep running in the background, and input owed on four other open roles. All the faster shortlist does is join that manager's queue sooner.

How long is the queue? One estimate for mid-market companies puts the gap between shortlist delivered and first interview scheduled at five to eight business days. Whatever the exact figure is where you work, it's the part of the process the new tools didn't touch.

People who run factories learned this a long time ago. Speeding up a step that isn't the slowest one doesn't get more out the door. When recruiter screening was the slow step, making it faster really did help. Now that it's fast, the slow step is the hiring manager's availability, and further gains in screening speed mostly produce shortlists that wait. If your time-to-shortlist chart looks great and your time-to-offer chart is flat, you've been measuring a step that is no longer the one that limits you.

The wait costs more than it seems to, in two ways that rarely get tracked.

One is what happens to the candidate. People apply when they're motivated, usually in the middle of an active search and in response to something that caught their interest. Eight days later that interest has cooled. Other conversations have moved forward, and your role has gone from something they were excited about to a process they happen to still be in. If the manager had looked at the shortlist the day it arrived and interviewed the next morning, they'd be talking to the same person with more energy and more of their attention on this particular job.

The other is the manager's hours themselves. Those three slots a week are scarce and expensive. When the screening before them was shallow, they get spent on people who looked fine on paper but were never a real fit. You might get one good conversation and two obvious mismatches that should have been caught weeks earlier. When screening is thorough, all three slots can go to people worth talking to.

That second cost points to the way out. Most AI hiring tools are built for the recruiter, which makes sense, since that's where the volume is. But the more useful question now is how better screening changes what a hiring manager gets from each hour of interviewing.

Here is a hypothetical. Say a manager usually needs to interview eight people to make one hire. If better screening got that down to four, they could hire twice as fast with exactly the same calendar. Nobody found them more time. They just stopped spending it on candidates who were never going to get an offer.

So I think the job of AI in hiring has shifted. Speed still matters, because it reaches candidates before their attention moves on. But once screening is fast, what moves time-to-offer is how good the screening is, which you can measure by how many of the manager's interviews turn out to be worth having.

This is the part of the problem Asendia was designed around. Plenty of fast screening is still shallow: it ranks resumes quickly and passes along people who look fine on paper, which is how a manager ends up with two obvious mismatches in three slots. Asendia calls each applicant soon after they apply, often that same evening, and holds a spoken screening conversation that adapts to their answers. A conversation shows things a resume doesn't, like how someone communicates, whether the specifics of the role fit, and whether they actually want this job. So what reaches the hiring manager is a structured summary of each conversation, with the candidate's own words quoted and the shortlist ranked, in place of a resume score, and it's written into the ATS where the rest of the candidate's record lives. Agencies running high-volume campaigns use it so the hiring manager gets involved once there's real information about each candidate, and not while there's still a raw pile of applications.

If you want to know whether this applies to you, put your time-to-shortlist and time-to-offer side by side for the last few quarters. If the first fell and the second didn't, the constraint has moved. And if you're rethinking which numbers to watch at all, the post on recruitment KPIs in a post-AI world is a good place to start, particularly on why offer-acceptance rate is a lagging measure of how good your pipeline really was.

Ready to transform your hiring strategy? Schedule a Demo with our founders today!