Let’s start with an uncomfortable truth.
Most hiring decisions used to be made the same way we choose what to watch on Netflix: a mix of instinct, familiarity, and questionable judgment.
And for years, we got away with it.
Until someone finally did the maths.
A bad hire can cost between 50% and 150% of annual salary depending on seniority, factoring in lost productivity, rehiring, and team disruption.
Suddenly, “gut feel” doesn’t feel so charming anymore. So we brought in AI.
🤖 The AI Gold Rush (Or: “This Will Fix Everything”… Right?)
AI in hiring is now used by over 50% of organisations globally to screen, assess, and shortlist candidates.
(World Journal of Advanced Research and Reviews, 2024)
On paper, it’s brilliant:
IBM reported 25% lower first-year attrition using predictive analytics
Unilever saw ~16% improvements in retention
Time-to-hire reductions of up to 35% have been widely cited
If you’re a hiring manager, that sounds like magic.
If you’re a recruiter… it sounds like your job just got automated.
Except, it didn’t.
🧪 The Bit We Don’t Talk About Enough: Is This Science or Sales?
Here’s where things get murky.
Many AI hiring tools claim to be “predictive.”
But predictive based on what?
Some are trained on historical hiring data (which may reflect past biases)
Others rely on behavioural proxies that sound scientific but lack transparent validation
Very few vendors openly publish long-term, peer-reviewed evidence
(The Algorithmic Turn in Talent Acquisition IRE Journals Mar25)
Yes, structured behavioural interviews are strongly validated - so much so that:
It takes four unstructured interviews to match the reliability of one structured one
They’ve been shown to reduce turnover dramatically (e.g. ~39% → ~13%)
(BECKY J. OLIPHANT, Stetson University)
But not all AI tools are built on this level of rigour.
Some are closer to “slick dashboards with confidence.”
⚖️ Bias: Removed, Reduced… or Repackaged?
AI is often sold as the antidote to human bias.
And to be fair, structured, standardised processes do reduce variability and subjectivity.
But here’s the catch:
AI doesn’t eliminate bias.
It scales whatever it’s trained on.
If historical hiring favoured certain schools, backgrounds, or personality types… AI can quietly learn that pattern and optimise for it.
The difference?
Human bias is visible (and challengeable)
Algorithmic bias is hidden (and scalable)
That’s a very different risk profile.
💸 Cheaper, Faster… Better?
One of the biggest shifts we’re seeing at TENTEN Partners is the democratisation of hiring technology.
What used to cost enterprise budgets is now:
Subscription-based
Plug-and-play
Available to almost any hiring team
Skills tests. Behavioural assessments. AI screening tools.
All cheaper. All faster. All promising better outcomes.
And in many cases, they do help. Scenario-based and behavioural assessments can predict performance with up to ~80% accuracy when combined with structured data.
But there’s a subtle danger here: Not everything that matters in hiring shows up in a dataset.
👀 So… Are Recruiters Becoming Obsolete?
Short answer: no.
Less comforting answer: the role is being exposed.
If a recruiter’s value is:
CV screening
Keyword matching
Basic coordination
Then yes - AI will replace you.
But the real value of recruitment was never supposed to be that.
It’s:
Challenging hiring managers who want “more of the same”
Interpreting ambiguous human potential
Balancing data with context
Spotting what doesn’t show up in a CV or algorithm
In a world of abundant data, judgment becomes more valuable, not less.
🧠 The Shift That Actually Matters
The most effective organisations today aren’t asking:
“How do we automate hiring?”
They’re asking:
“How do we make better decisions—with better tools?”
That’s why we’re seeing a move toward:
Skills over credentials
“Culture add” over “culture fit”
Hybrid models where AI handles scale, and humans handle nuance
Because hiring isn’t just a matching problem.
It’s a prediction problem.
A behavioural problem.
And occasionally… a leap-of-faith problem.
🚀 Final Thought (No Buzzwords, Promise)
AI will absolutely change recruitment.
It will make it faster, more data-driven, and, at its best - more fair.
But it won’t remove the hardest part of hiring:
Deciding which imperfect human is most likely to succeed in an imperfect job, in an unpredictable future.
No algorithm has cracked that yet.
And until it does, we’re all still in business.
At TENTEN Partners, we work with clients across financial services and technology to navigate exactly this balance—between data, judgment, and what actually drives long-term performance.
If hiring is becoming more scientific, the real question is: are we applying the science properly?
