Reskill or die: inside a Nordic IT academy

The consulting industry spent the last decade selling transformation. In 2026, it's the thing being transformed.

IDC puts the cost of the IT skills gap at $5.5 trillion by year-end. Robert Half's data says 7% of tech leaders think their team has the skills to deliver their top projects. And the firms that built their business on closing skills gaps for clients are now laying people off to afford the upskilling bill for their own workforce.

That's the tension. I run sales at a Nordic IT academy that reskills consultants for a living, so I watch this from a specific angle. Here's what the numbers actually say, and what most operators are getting wrong.

Key takeaways:

• 89% of leaders say AI skills are critical. Only 6% have started meaningful reskilling (MetaIntro, 2026). That's not a gap, it's a canyon.
• McKinsey runs ~20,000 internal AI agents, up 500% in 18 months, and cut ~10% of its workforce. Deloitte scraps its traditional job-title ladder on June 1, 2026. Big 4 UK graduate listings are down 44% year over year.
• The firms doing the loudest reskilling are also the ones restructuring hardest. Accenture committed $3B to data and AI plus LearnVantage. KPMG committed $2B. Infosys has certified 270,000 of 300,000 employees as "AI Aware."
• Nordic numbers are worse than the headline: 66-90% of Finnish firms can't find the skills they need, and the region produces only ~5,000 engineers a year across all four countries.
• Experis Academy in Norway runs at a 96% placement rate. The academy is the moat, not the marketing.

The numbers are ugly and almost nobody is moving

89% of leaders call AI skills critical. 6% have meaningful reskilling in motion. That gap isn't an HR problem. It's a strategy problem wearing an HR badge.

IDC expects the skills shortage to cost organisations $5.5 trillion by the end of 2026 in product delays, lost business, and competitive erosion. Robert Half's latest tech hiring survey says only 7% of tech leaders feel their team is equipped to deliver their priority projects. Gloat puts the AI-skills wage premium at 56%, up from 25% the year before.

Demand is compounding. Supply isn't. Everyone knows this and almost nobody is acting on it.

Here's the uncomfortable part. A lot of firms made big reskilling announcements in 2023 and 2024, then quietly shelved the spend when the AI hype cooled and margins got tight. The 6% number is what's left once you strip out the PR.

The Big 4 are restructuring, not adapting

McKinsey's own tooling tells the story. Bob Sternfels said internal AI agents at the firm grew 500% in 18 months to roughly 20,000. The firm cut ~10% of its workforce and automated 200 internal tech and support roles. Fast Company reads this as a warning for consulting. I read it the same way.

Deloitte is scrapping its traditional US job titles on June 1, 2026. Fortune framed it as a "modernization." Translation: the pyramid that consulting was built on, where juniors did the grunt work and partners billed the insight, is collapsing because AI does the grunt work for free.

The graduate pipeline shows it first. Big 4 UK graduate job listings are down 44% year over year. EY has delayed its graduate start dates three years in a row. The entry-level tier of the consulting industry is being quietly deleted.

This isn't visionary. It's defensive. The Big 4 aren't reskilling because they believe in lifelong learning. They're reskilling because their own operating model stopped compounding.

The firms that are winning built internal academies

Look at who's actually spending:

• Accenture: $3B into data and AI, plus LearnVantage, a learning platform running through nano-degrees and Stanford partnerships. ~70,000 employees trained on agentic AI as of early 2026.
• KPMG: $2B into AI, with an AI-driven LMS embedded inside the workflow so upskilling happens while people work, not in a classroom on Tuesdays.
• Infosys: 270,000 of ~300,000 employees certified "AI Aware." Klover's analysis calls this a "human capital moat." That's the right framing.
• EY: $1.4B over five years. Slower to deploy but spending at the same order of magnitude.

The academy is the moat. Every one of these firms is monetising its internal upskilling programme as a client offering. You can't sell AI transformation if your own consultants can't spell agentic.

Meanwhile, CareerMetrics reports that 1 in 3 firms that laid off for AI are already rehiring. The replacement-by-automation bet didn't work. The reskilling bet did. That's the split.

The Nordic market breaks harder than the headlines

Now the local view. European consulting narratives get written from London and Frankfurt. The Nordics run on a different math.

• Finland: 66-90% of firms say they can't find the skills they need, depending on industry (Grid Dynamics).
• Sweden: 56-65% of enterprises report a shortage of seasoned tech professionals.
• Nordic countries combined produce roughly 5,000 engineers per year. The education system cannot keep pace with the market.

The macro argument says "upskill or import." In the Nordics, both levers are constrained. Immigration to fill tech roles is politically slow. Domestic pipelines are thin. The only thing a Nordic consultancy can actually control is its internal reskilling programme.

Which is exactly why the firms that treat the academy as a P&L line, not an HR cost centre, are pulling ahead.

What I'm watching from inside Experis Academy

I've been Head of Sales at Experis Academy in Oslo since 2024. We reskill tech consultants for a living. Here's what I see that doesn't show up in the McKinsey decks.

We sit at the pointy end of the Nordic IT consulting market. Clients come in with two kinds of problem. One, their own internal team can't ship what the business is asking for. Two, they've hired a consultancy and the consultancy can't ship what the business is asking for either. Both problems trace back to the same root: the skill mix lags the demand curve by 18-24 months, and no traditional learning pipeline closes a 24-month gap in a quarter.

What works:

• Outcome-based cohorts, not generic courses. Structured reskilling tied to real client projects, with a deliverable at the end, transfers skill. A course library on its own doesn't.
• Reskilling as a client offering, not a benefit. The same programme that reskills your own consultants becomes the product you sell to enterprise clients who need to reskill their workforce. Same muscle, two revenue lines.
• Placement discipline. Experis Academy runs at a 96% placement rate into IT roles. That's the number that matters. Training is a cost. Placement is the return.

What doesn't work:

• Buying licenses to an e-learning platform and calling it a reskilling strategy. No cohort, no project, no outcome = no transfer.
• Training only the top 20%. The reskilling premium comes from moving the middle. The stars will learn anyway.
• Treating AI fluency as a one-time event. Infosys's "AI Aware" framing is smart because it implies a floor, not a ceiling.

The firms that are going to lose the next three years are the ones who keep thinking of reskilling as a project with a budget and an end date.

One more observation from the Nordic context. The supplier landscape here is consolidating. Smaller boutique consultancies that relied on hiring senior specialists off the open market are getting priced out as specialist compensation climbs faster than billable rates. Mid-sized firms that build their own academies are the ones still winning tenders. The math is brutal but clean. If you can't buy the skill and you can't build it, you lose the client.

The only consulting strategy that survives 2026

Strip out the PowerPoint and the operating model is simple. If you run a consulting business, a consultancy practice inside a bigger firm, or a team that sells expertise:

1. Treat the academy as infrastructure, not overhead. Accenture, KPMG, Infosys aren't spending billions because they're generous. They're spending because it's the only asset that still compounds at AI speed.
2. Shrink the pyramid deliberately. The grad-heavy pyramid is dead. Design for fewer, better-leveraged consultants, each augmented by tools that used to require two juniors.
3. Sell the reskilling, don't just do it. If you can move your own people, you can move the client's. That's a product, not a perk. The same logic holds at the individual level: see why AI tools are not a strategy and how to learn AI by building.
4. Measure placement, not training hours. Hours trained is a vanity metric. Projects shipped, roles filled, revenue per consultant are the ones that matter.

The $5.5T number is scary. The 6% number is the real signal. Most firms still treat this like an HR initiative. The ones that treat it like a category-defining bet are the ones who'll still be here in 2028.

I'm watching this from inside one of them. The work is unglamorous. The maths is obvious. Reskill or die isn't a slogan. It's a balance sheet.