The Price of Intelligence Just Collapsed. Most Businesses Are Still Paying the Old Price.

Alan Kohler published something last weekend that I haven't been able to stop thinking about.

He made a single, brutal observation: the AI revolution is, at its core, a collapse in the price of intelligence.

Large language models price their output by the token, roughly a million tokens per 750,000 words. A few years ago, querying an AI at GPT-3.5 performance level cost around $20 per million tokens. Today, the same performance costs less than $0.07. That's a 280-fold reduction in three years, and it's still accelerating. According to Stanford's 2025 AI Index Report, LLM inference costs are falling at a median rate of 50 times per year at any given performance level. Bloomberg called it the sweet part of the S-curve  just this week.

Kohler compared this to what freelance writers earn. The going rate in Australia is roughly 50 cents per word if you're lucky. That same 750,000 words costs $375,000 in human labour — and would take a writer producing two 1,000-word articles per week about eight years to produce. AI generates it for roughly $35 Australian. A 99.99% discount.

His conclusion was direct and uncomfortable: every white-collar job is facing the same commoditisation of intelligence that journalists are facing right now.

And then he said the thing that stopped me:

"We are the incumbents."

He's right. And I'd go further.

We are the incumbents. And most of us are still acting like it's 2019.

The problem isn't that businesses don't know AI is coming. Everyone knows. The problem is that most organisations know AI is disrupting other industries and still believe their hiring process is somehow exempt.

It isn't.

Hiring is one of the most intelligence-intensive, labour-heavy processes in any business. Think about what actually happens when a role goes live. A hiring manager spends time writing a job description. The role gets advertised. 150 to 300 applications come in. Someone sifts through CVs for hours. Phone screens get booked. First-round interviews happen. Notes get lost. Great candidates drop out because the process takes too long.

According to AMS and the Josh Bersin Company, the global average time-to-hire is now 44 days, up from 43 days a year ago. It's getting worse, not better. SHRM data puts the average cost per hire at nearly $4,700 in the US alone, not including the hidden cost of the role sitting vacant, which Deloitte estimates at around $500 per day.

And if you're using an agency? Industry-standard fees run 15–25% of the first-year salary. For a $120,000 role, that's $18,000 to $30,000 per placement. Per hire.

Nobody questions this. It's just absorbed as the cost of hiring.

It's not. It's the cost of not adapting.

I built Outstaffer because I saw this coming.

I want to be honest about the timeline here.

I founded Outstaffer.com in 2020. The original thesis wasn't about AI; it was simpler than that. The problem with global hiring wasn't that the talent didn't exist. It was that the HR infrastructure to find it, assess it compliantly, and hire it across borders was completely broken. Companies either set up expensive legal entities in each country, used outsourcing or agencies and paid massive fees, or misclassified contractors and hoped no one noticed.

None of those options were good enough.

But as we built the platform and watched AI develop, something became increasingly obvious: the most expensive, most broken part of the entire hiring funnel wasn't the legal infrastructure. It was the front end. The screening. The repetitive, high-volume, cognitive work that sits between posting a job and shortlisting a candidate.

That work (reading hundreds of CVs, running first-round interviews, scoring candidates against a rubric, ranking a shortlist) is precisely the type of intelligence-heavy, volume-dependent task that AI can do faster, cheaper, and more consistently than any human.

Not better than a great recruiter making a nuanced, contextual judgment call. But better than a burnt-out hiring manager on their 47th CV of the week? Absolutely. And research backs this up.

A University of Chicago Booth study covering over 70,000 applications found that candidates who went through AI-led interviews succeeded in subsequent human interviews at a rate of 53% - compared to 28% from traditional resume screening. The same study found that 78% of applicants, when given the choice, preferred the AI interviewer over a human one. AI interviews resulted in 12% more job offers, 18% more job starters, and 16% higher 30-day retention rates versus traditional hiring.

These aren't marginal improvements. That's a different class of outcome.

So we built Anna.

What Anna actually does and what she doesn't.

Anna is Outstaffer's conversational AI. She conducts real, structured, human-like interviews with every applicant, 24 hours a day, across time zones. She assesses candidates across 100+ skill dimensions: communication ability, critical thinking, CEFR English language levels, and role-specific competencies. Every interview is recorded, transcribed, and scored.

Then our human recruiters step in. They review every result, validate the outcomes, select the best candidates and deliver ranked shortlists, typically within 4 to 7 days. You only meet AI-assessed, recruiter-validated finalists. No CV slog. No screening calls. Just meaningful final interviews with the people who are actually worth your time.

The result is 15 to 20 hours saved per role. Verified shortlists in under a week. And a candidate NPS of 90 to 98%, because every applicant gets an interview, an outcome, and a response. Most companies can't say that.

Here's the thing I want to be clear about, because this is where a lot of AI hiring products get it wrong: Anna automates 90% of the process. But Outstaffer recruiters own the outcome. Recruitment is too high-stakes, too nuanced, and too consequential for a purely automated system. The AI accelerates the process. Humans safeguard the decision. That distinction matters.

And critically, clients don't operate software. They receive outcomes.

But finding great talent is only half the problem.

Here's the part most people miss when they talk about AI and the future of hiring.

According to ManpowerGroup's 2026 Global Talent Shortage Survey (conducted across 39,000+ employers in 41 countries), 72% of employers globally are now reporting difficulty finding the skilled people they need. In Australia, that's 65% of employers. In the UK, 73%. In the US, 69%. In India, 82%.

Korn Ferry projects that by 2030, more than 85 million jobs globally could go unfilled, representing $8.5 trillion in unrealised annual revenue. For Australia, the UK, the US, and their peers, this isn't a distant forecast. It's an accelerating present.

The talent you need almost certainly exists. The problem is it doesn't live in your city, or your country, or your traditional hiring geography. In 2025, cross-border remote hiring grew 38% year-over-year. The best candidate for your next role might be in the Philippines, Vietnam, India, Malaysia, or Singapore. They're skilled, they're motivated, and they're available.

The barrier isn't finding them. We just solved that with Anna. The barrier is hiring them legally, without setting up a local entity in each country or using an outsourcer.

This is where Employer of Record closes the loop.

The full stack: AI finds the talent, EOR unlocks the hire.

Through Outstaffer's EOR service, once our AI Recruiter service has found and shortlisted the right candidate (in any of our supported markets), you can hire them compliantly within days.

We become the legal employer on record in the employee's country. We handle the employment contracts, payroll, tax, benefits, statutory entitlements, and ongoing compliance. You get the person. We handle everything else, including the things Outstaffer was built to make simple: onboarding, right-to-work checks, AI-powered identity verification, and ongoing HR management through our Borderless HRIS.

We currently operate across Australia, the Philippines, Vietnam, India, Singapore, Malaysia, and Thailand, with further markets planned.

That's the full stack. Outstaffer AI Recruiter finds and shortlists globally. EOR makes the hire legal and compliant. HRIS manages the employee lifecycle from day one.

Find. Hire. Manage. One platform.

The uncomfortable question every leader needs to answer.

Kohler's piece wasn't really about journalism. It was a warning about every industry that runs on human cognitive labour. And hiring (the entire talent acquisition function) runs almost entirely on human cognitive labour.

The incumbents who adapt will hire faster, spend less, and access a global talent pool their competitors can't reach. According to SHRM, 43% of organisations are now using AI for HR tasks, up from 26% just a year ago. Of those, 64% are using it specifically for recruiting, interviewing, and hiring. 89% say it saves significant time. The adoption curve is steep, and it's not slowing down.

The ones who don't adapt will keep paying agency fees of 15–25% per placement. Keep waiting 44 days to fill roles. Keep losing great candidates to businesses that move faster. And keep managing global talent through a patchwork of contractors, misclassification risks, and manual workarounds that don't scale.

That's not speculation. That's already the gap between the companies I speak to who've made this shift and the ones who haven't.

The price of intelligence collapsed. The price of finding and hiring intelligent people is collapsing right behind it.

We built Outstaffer because I believed that was coming. I still believe the companies that move now, not in 18 months, will have a structural hiring advantage that compounds over time.

The question isn't whether your hiring process needs to change.

The question is whether you're going to change it before your competitors force you to.

Outstaffer is an All-in-One AI Recruitment, HR, and Employer of Record platform. We help businesses find, hire, and manage talent locally and globally — faster, fairer, and at a fraction of the traditional cost.

Ready to see it in action? Try your first role free‍