The "Yes" Problem: Why Most Screening Tools Are Built Backwards

There is an assumption baked into almost every CV screening tool on the market. It goes largely unexamined because it feels so obviously correct that nobody thinks to question it.

The assumption is this: the purpose of screening is to find reasons to say no.

Build a list of requirements. Match candidates against it. Filter out anyone who doesn't hit the threshold. What remains is your shortlist.

It sounds logical. It is also, in practice, one of the most reliable ways to miss good candidates.

The filter problem

Screening built around elimination is only as good as the criteria it eliminates against. And recruitment criteria, more often than not, is written by people describing what their last good hire looked like rather than what their next good hire needs to be.

The result is a filter that's optimised for the past. It catches people who resemble previous success and removes people who don't - regardless of whether those people would actually do the job well.

Add in the inconsistency of how candidates describe themselves - one person writes "led a team of eight", another writes "team leadership across multiple functions" - and a keyword-based filter doesn't just miss bad candidates. It misses good ones. Systematically, and invisibly.

What "finding the yes" means

The alternative isn't lowering standards. It's changing the question.

Instead of "does this candidate fail any of our criteria?" the question becomes "what is the strongest case for this candidate, and does that case hold up?"

It's a small shift in framing but a significant shift in outcome. A recruiter approaching a CV looking for reasons to progress it will read it differently from one looking for reasons to reject it. The same CV, the same words, produces a different verdict depending on which question is being asked.

The best recruiters have always known this intuitively. The ones who build the deepest pipelines are the ones who find the yes in a CV that a less experienced colleague would have filed away. They read between the lines. They ask the follow-up question. They make the call.

The problem is that this skill doesn't scale. One experienced recruiter can do it across twenty CVs. They cannot do it across two hundred.

Where AI gets this wrong

Most AI screening tools haven't fixed this. They've just automated the filter. The machine now does the elimination faster and at greater scale - which means it also misses good candidates faster and at greater scale.

The issue isn't the technology. It's the design philosophy. If you build a tool to find reasons to say no, and then you power it with AI, you get a very efficient reason-finding machine. You do not get a better recruiter.

A better recruiter - human or AI - approaches each candidate with a different orientation. What's the strongest version of this person's story? What context am I missing? What question would change my view?

The clarification moment

The most interesting thing a screening tool can do isn't produce a verdict faster. It's identify the moment where a question would change the outcome.

The candidate whose CV looks thin because they haven't described their experience well. The applicant who looks overqualified on paper but has a clear reason for the move. The person whose background is unconventional but whose transferable skills are exactly what the role needs.

In each of these cases, the difference between a yes and a no isn't the CV. It's one conversation. A single question asked at the right moment that turns an uncertain shortlist decision into a confident one.

That's the "yes" problem. Not that recruiters don't want to find the right candidates. They do. It's that the tools they use aren't designed to help them.