The Job Search Has Changed. And We're Optimizing for the Wrong Thing

Finding a job today is no longer just about having the right experience:

• Your CV gets uploaded.

• Then an ATS scans it.

• AI looks for keywords.

• Another system may rank your profile.

And somewhere along the way, a perfectly qualified candidate can disappear into a rejection email without anyone ever knowing they applied.

But let's look at both sides.

Imagine you're a recruiter and 500 people apply for one position.

You can't realistically read 500 CVs in detail.

So you start filtering.

• Years of experience

• Job titles

• Keywords

• Certifications

• Location

• Salary expectations

• Industry experience

• Education

• AI-generated candidate scores

Maybe 500 applications become 150. Then 150 become 50. Then 50 become 10.

The problem is that somewhere between 500 and 10, the person who could have been perfect for the job might have disappeared. And the recruiter may never know it happened.

Now look at the candidate's side.

Job seekers have figured out how the system works too.

They're being told to:

• Optimize their CV for ATS systems

• Copy keywords directly from the job description

• Generate cover letters with AI

• Rewrite their experience for every application

• Match the exact terminology used in the vacancy

• Sometimes even manipulate formatting to get past automated screening

I've seen people discuss putting keywords into CVs in tiny fonts or hiding them in white text just to satisfy automated filters. That's where things get ridiculous.

We're now teaching people how to write documents that machines like instead of documents that accurately represent the person and AI-generated applications create another problem.

If everyone uses AI to produce the "perfect" cover letter, eventually every cover letter starts sounding perfect.

• Professional.

• Polished.

• Generic.

And almost completely indistinguishable from the next one.

So who actually benefits?

1. The recruiter needs a way to deal with volume.

2. The candidate needs a way to get noticed.

3. The ATS needs structured data.

4. And AI needs something it can score.

5. Everyone is optimizing their part of the process.

But the result can be a hiring system where everyone is optimizing and nobody is actually talking to each other.

That's the part I find interesting, I'm not against AI in recruitment quite the opposite.

AI can be extremely useful for:

• Removing repetitive administrative work

• Summarizing CVs

• Matching experience to requirements

• Scheduling interviews

• Finding relevant candidates

• Identifying gaps in a candidate's experience

The problem starts when screening becomes the hiring process instead of helping the hiring process.

What would a better approach look like?

It doesn't necessarily mean throwing away ATS systems or making recruiters manually read 500 CVs. It could be something much simpler.

Instead of asking:

"Does this CV contain the right keywords?"

Ask:

"Is there enough evidence here that this person deserves a conversation?"

That's a very different question, a 10-minute screening conversation can sometimes tell you more than another round of keyword matching.

A short practical task can tell you how someone actually works.

A portfolio can show what someone has built.

A structured interview can reveal experience that a CV simply doesn't communicate and AI could still help with all of this.

Use AI to reduce the workload, not to remove the human being from the equation.

And here's the uncomfortable part.

Companies often say:

"We can't find good people."

Candidates often say:

"I can't even get an interview."

Both statements can be true at the same time there may be plenty of qualified people, there may be plenty of open positions.

And yet the two sides can still fail to connect maybe that's not a talent shortage maybe it's a connection problem.

One thing companies could change tomorrow

If you're a hiring manager or business owner and you have 500 CVs waiting for you tomorrow morning, don't throw away your filters.

• But don't let the filters make the final decision either.

• Take the candidates who are close to matching.

• Not just the ones who perfectly match the keywords.

Then give a small number of them a structured 10-minute conversation or practical screening task you might discover that the person with the "wrong" job title has exactly the experience you actually need.

Because a CV is not a person, a keyword is not competence and an AI score is not an interview.

The goal of recruitment shouldn't be to find the candidate who is best at getting through the recruitment system. It should be to find the person who can actually do the job.

Maybe it's time we started optimizing for that again.