The role of recruiting and talent acquisition
How pre-qualification and recruiter interviews work together to build your shortlist in the age of AI.
Every open role today starts the same way: a flood of applicants, and a hard question. How do you get from 300 applications down to the 10 best candidates to put in front of the hiring team? That job, narrowing the flood and delivering a high-quality candidate shortlist, is what the recruiting and talent acquisition team owns. The answer is a series of stages, each one narrower than the last. Some of those stages are safe to automate. One of them, the most important one, belongs to a recruiter. Here’s how the whole funnel works, and where the human still wins.
300 job applicants: the flood
Applying to jobs used to take effort. Now it takes one click. Job seekers use AI tools to fire off resumes to hundreds or even thousands of openings at once, often in under a minute each.
The result for the employer is a wall of applications. The average recruiter today receives roughly four times as many applications as they did just a few years ago. A single role can pull in 300 applications, sometimes 150 on the first day it’s posted. And many of those resumes were written by AI, so they look polished and similar, which makes it even harder to tell who’s actually good.
No human can read 300 resumes carefully, let alone talk to everyone. So the funnel has to do the work: narrow 300 down to a shortlist of 10, without losing the good people along the way.
This idea of a hiring funnel isn’t new. First Round Review’s classic piece, The Simple Numbers That Could Change How You Hire, lays out the “divide by four” rule: at each stage, roughly a quarter of candidates should move forward, from sourcing to screening to interviewing to hire. That math still holds. What’s changed is the sheer volume pouring into the top, and the new AI tools now crowding into the middle.
50 are qualified: pre-qualification
The first stage is pre-qualification. It asks simple yes/no questions to find out who even meets the job’s basic requirements. These are the knockout questions you see on an application:
- Are you authorized to work in this country?
- Do you have the required license or certification?
- Do you have the minimum years of experience?
- Are you willing to work on-site, if the role requires it?
Everyone gets the same questions. The answers are pass or fail. Nobody is judged or scored; the stage just checks a fact and lets the qualified people through.
Here’s the useful part: this stage is adjustable. If your knockout questions are loose, maybe 150 of your 300 applicants “meet the minimum.” That’s still far too many. So you can tighten the questions and raise the bar, for example, require a specific certification or a higher experience minimum, and cut the field down to around 50 truly qualified people.
Because this stage is just checking facts with the same rules for everyone, it’s safe, fair, and easy to automate. Software handles it well. This is what “pre-qualification” should mean: a yes/no minimum-bar check, nothing more.
But qualifying isn’t the same as being one of the best. You can’t send 50 people to a hiring manager. So how do you go from 50 qualified applicants to the 10 best? This is the hard part of hiring, and it’s where the real disagreement is happening right now.
30 reviewed: the AI pre-screener
A growing number of AI tools try to narrow that qualified pile for you. They ask candidates to record video or voice answers, then the software grades those answers and ranks the candidates against each other.
The pitch sounds nice: “Let every candidate show more than their resume.” And it’s a bit more interactive than a yes/no application question. Used carefully, an AI pre-screener can help thin a large pool of qualified candidates, say from 50 to 30, before a recruiter spends live time.
But notice what’s happening when these tools score and rank: this isn’t a fact check anymore. It’s a machine making a judgment about a person, and ranking them, before any human has talked to them. That’s a much riskier place to be, and the law has started to catch up. A few examples as of 2026:
- New York City’s Local Law 144 requires an annual independent bias audit of automated tools that score, rank, or classify candidates, along with publicly posted results and advance notice to candidates. A December 2025 audit by the New York State Comptroller found the city’s enforcement of the law has been falling short, a sign this area is only getting more scrutiny.
- Illinois requires employers to tell candidates when AI will analyze their video interview, explain how it works, and get consent first. Companies have already been fined for skipping this.
- Colorado and other states are rolling out broader rules for AI that influences hiring decisions.
There’s a quality problem too. About 1 in 5 companies that use AI in hiring say their tools have screened out qualified candidates by mistake. And trust is low: in one survey, only 8% of job seekers believed AI screening of applications makes hiring fairer.
The lesson isn’t “AI is bad.” It’s that a yes/no minimum check is safe to automate, but judging and ranking real people is a different, riskier job, and it’s exactly the job a recruiter should own. Which is why the next stage is the one that matters most.
20 interviewed: recruiter interviews
This is the stage where a recruiter actually talks to the candidates. It’s the live conversation in which a recruiter evaluates qualified candidates and decides who’s truly worth the hiring team’s time.
For years, this got called a “screening call,” and that name sold it short. It was never just screening. It’s the first interview, and it’s the most valuable step in the whole process. We’ve long argued that candidate screening is the most overlooked step in recruitment, the stage with the most room for improvement. That was true before AI pre-screeners arrived. It’s even more true now that recruiters can hand the yes/no checks to automation and focus their energy on the conversation itself.
Here’s the key difference between this stage and pre-qualification. Pre-qualification is a setting you can tighten or loosen, a yes/no threshold. The interview is not a setting. You can’t turn a dial to know who’s truly the best out of a qualified pool. A person has to evaluate them, ask real questions, and make a judgment call. That’s why this stage belongs to a human.
And this is the work that’s irreplaceable. When recruiters conduct skills-based and competency-based interviews on the first call with a candidate, recruiters are irreplaceable. A machine can check a box. It can’t have this conversation.
When the earlier stages handle the minimum-bar facts, the recruiter doesn’t waste the call re-confirming work authorization or salary range. That frees them to run a real, job-specific first interview, using a call guide built for the role, asking the skills and competency questions that reveal who’s actually strong. The recruiter moves from “person who confirms the basics” to “person who finds the best.” No AI tool can fill that role.
It also makes hiring faster. A strong first interview surfaces the real picture of a candidate up front, so the hiring team isn’t burning week after week on people who looked fine on paper but fall apart in later rounds. Better signal early means fewer wasted interviews later. That can take weeks out of the process.
10 top candidates: the shortlist
At the end of the funnel is the candidate shortlist: the 10 or so people worth presenting to the hiring team. This is the payoff, and it’s the part only a recruiter can build.
Meeting the minimum is not the same as being one of the best. A yes/no check can tell you who qualifies. An AI tool can help thin a large pool. But only a recruiter’s interview can decide who actually makes the shortlist and who to champion.
To be clear about where the handoff happens: the recruiting and TA team’s job is to deliver this shortlist, not to pick the hire. From here, the candidates move on to the downstream process the hiring team owns: the on-site rounds, the panel interviews, the conversations with hiring managers, executives, and other stakeholders who make the final call. That’s exactly the point. When recruiters deliver a shortlist of genuinely strong, well-vetted candidates, every downstream interview starts from a better place, and the hiring team can focus on choosing among good options rather than weeding out weak ones.
How it all works together
Here’s the whole funnel in plain terms:
- 300 applicants: the flood, many applying in one click with AI-written resumes.
- 50 qualified, pre-qualification: yes/no knockout questions confirm who meets the minimum. Adjustable and safe to automate.
- 30 reviewed, AI pre-screener: optional, can help thin a large qualified pool, but scoring and ranking people carries real legal and fairness risk.
- 20 interviewed, recruiter interviews: the human evaluation, where a recruiter has the conversation a machine can’t.
- 10 top candidates, the candidate shortlist: the best of the best, delivered to the hiring team.
From there, the hiring team takes over: the on-site rounds, panel interviews, and final decision belong to the hiring managers, executives, and stakeholders. The recruiter’s job is to hand them a shortlist worth their time.
Used in the right order, these stages don’t compete. The automated checks screen out people who were never eligible, so the recruiter can spend their time on high-quality interviews with candidates who actually deserve a real conversation. The smartest hiring teams let the simple checks handle the minimum bar, and let recruiters do what they do best: talk to people and find the best ones to put in front of the hiring team.
Honeit is built around the recruiter’s first interview, with job-specific call guides, automatic note-taking, scorecards, and one-click candidate presentations that share the candidate’s actual answers. (And yes, a Honeit pre-qualification voice tool is coming soon, to handle the yes/no minimum-bar checks before the conversation begins.)
