AI-native, work-ready, and nobody's hiring them
Also published on LinkedIn
There's a decision being made in budget meetings right now. Don't backfill the juniors. AI does that work now. We hear the same sentence in rooms every week, and Gartner puts numbers on it: one in five CHROs already have a business leader who has frozen entry-level hiring because of AI.
And while they freeze, some of the biggest operators are moving hard the other way.
IBM is tripling its entry-level hiring in the US this year. Not despite AI, because of it. "And yes, it's for all these jobs that we're being told AI can do," is how people chief Nickle LaMoreaux put it. Singapore has made the same bet national strategy, putting over a billion Budget dollars behind tripling its AI workforce.
Same technology, opposite conclusions. One side sees a cost line. The other sees the best-value talent in the market.
The fastest adopters money can buy
The hardest problem in every AI transformation was never the tooling. It's adoption. We see the fluency gap everywhere: a handful of people compounding gains weekly while the majority circle the tools. Culture Amp's people chief Justin Angsuwat made most of his workforce confident with AI and watched behaviour barely move: people hit their Tuesday morning workload and revert. His most senior people often struggled most; unlearning twenty years of habit is harder than learning fresh.
Graduates solve that problem on arrival. Atlassian found new grads are 1.5 times more likely to use AI daily than the rest of the workforce. AI-native, no old workflow to defend, pulling everyone around them up the curve. Every AI programme budgets a fortune for change management. Graduates are change management that shows up wanting a job.
One catch: momentum only counts if you can see it. Track adoption role by role and your graduates become the leading edge of the rollout.
These aren't the old entry-level roles
Graduates bring AI fluency. The business adds context: whether the output is right, why the exception exists, what the customer actually meant. Context is built one way, doing real work next to people who have it. That's what the bottom rung was always for. The apprenticeship.
The winning move isn't hiring the 2019 graduate role back. That role is gone; its task list was mostly the work AI now does. The move is designing its replacement, and the replacements are already appearing. Roles are bundles of tasks, and AI has changed the bundle: strip out what the agents handle and what's left reaches upward. Work that sat a level or two up the ladder, client conversations, judgement calls, steering agent output, now fits inside a grad role because AI does the grunt underneath it. A grad with AI tooling is doing what mid-tier did three years ago. That's not a diluted junior role. It's a promoted one.
Build the muscle, don't rent it
The same maths settles the "AI specialist" question. Salaries for anything with AI in the title have surged, and most of it is overpriced: a specialist arrives with fluency and none of your context, the same half-equation the graduates have, at several times the cost.
We live this debate ourselves. Human Layer Lab is a fast-growing AI-native company: our workforce is human and agent together, every person runs a team of agents, and every role is designed at task level on our own platform, with a forward view of what it looks like in six months. And when we hire, the maths leans hard towards AI-native over experienced. Our context is young and quick to hand over, so fluency wins almost every time.
Most organisations are the opposite, and that's the point. You have an established workforce with years of context already on the payroll, the very thing we have to build one hire at a time. For you it isn't fluency or experience. It's both: bring in the AI-native juniors, pair them with the veterans who hold the context, and let them rise together, fluency handed up, judgement handed down. Then build what they both stand on: the tooling, the tasks and the processes, redesigned around an AI-first workforce of humans and agents working as one.
Singapore is running the play at national scale
Singapore isn't debating whether early-career AI talent is worth building. It's manufacturing it. AI Singapore's Apprenticeship Programme is fully funded with a stipend: deep skilling, then months embedded in real industry projects. Not coursework, production work. Most graduates land AI roles within months, into DBS, GovTech, Dyson and TikTok. The stated goal: AI builders, not certified users.
The message for employers: work-ready, AI-native junior operators are coming off these lines right now, some state-subsidised. Plug in early and get first pick. This is not a talent gap. It's a talent window.
The window is open
The question was never whether AI can do the old junior work. It can. The question is what your juniors do instead, and the answer is better work, sooner: in front of customers, making judgement calls, supervising the agents.
The pipeline maths is the same for everyone. Cut junior hiring and you're not saving, you're borrowing: from your own mid-level bench three years out, repaid at external-hire prices for people with none of your context. The firms hiring at the bottom now will own the mid-level market in 2029 while everyone else bids for what's left.
The most AI-fluent generation in history is entering the workforce, and a chunk of the market has frozen its intake. For everyone still hiring, that's the opportunity of the decade.
Sources: Gartner 2025-26 CHRO and manager research; IBM entry-level hiring announcement and Charter Leading With AI Summit remarks, February 2026; Atlassian Teamwork Lab, State of Teams 2026; AI Singapore AI Apprenticeship Programme; Singapore Budget 2026; Justin Angsuwat, Culture Amp, People Managing People interview, 2026.