AI Adoption Is Soaring... But...

We have a strange problem with AI. Everyone is using it. Apparently…

And yet almost no one is getting A LOT of value from it.

The data is stark. According to a global study by the University of Melbourne and KPMG, covering 48,000 people across 47 countries, 66% of people are now regularly using AI — but only 46% actually trust it.

57% rely on AI outputs without checking if they’re accurate. Which points to a lack of AI training.

And according to PwC’s 2026 Global CEO Survey — the largest of its kind, covering 4,454 CEOs across 95 countries :

Think about that. AI is everywhere. But impact, according to the CEO’s of major companies and organisations is…. nowhere.

We are living through the fastest adoption of a technology in history.

And one of the weakest returns on investment.

This isn’t a technology problem.

It’s a thinking problem.

The illusion of adoption

Right now, most organisations are doing the same thing.

• Buying licences.

• Running pilots.

• Encouraging “experimentation.”

But they are not changing how decisions get made.

So what happens?

AI gets used quietly, individually, and inconsistently. Not strategically. Not structurally.

Which is why, according to WalkMe’s 2025 research,

Shadow AI — invisible, unmanaged, and building no organisational value. Also it seems to be getting higher and higher as a percentage — with the earlier Personio report suggesting it was 67%… The key thing is…

If your employees are hiding AI use, your strategy has already failed.

AI is already breaking business models

If you think this is theoretical, it’s not.

Look at Chegg. Once a $14.5 billion edtech company. A market leader. Profitable and growing. Then ChatGPT arrived. Their core product — paid academic answers — became effectively free overnight.

Revenue fell 30% year-on-year. Then 49%. Market cap collapsed to around $100 million. Over 600 employees lost their jobs in six months. The NYSE has issued a delisting warning.

That’s a 99% destruction of shareholder value in under four years.

This is the Fifth Industrial Revolution. No one said it wouldn’t be bloody…

This is how AI disrupts. Not slowly. But suddenly.

AI doesn’t improve markets.

It deletes them.

So I decided to test something different. I’ve been talking for a while about the next phase of AI. Not rented intelligence. Owned intelligence.

So I tested a system built on that principle — MatrixStrike, a specialised intelligence tool built by Martin Lucas and his team at Gap in the Matrix.

Not on something easy. On something I know well. I’m speaking in June on the future of AgriTech, AI, and the Fifth Industrial Revolution.

It’s a genuinely complex space — global markets, climate risk, competing regulatory frameworks across the EU, India, and Brazil, and a funding correction that has wiped out 40–50% of sector investment since the 2021 peak.

So I asked the system to analyse it. I expected decent output.

What I got was something else entirely.

Not answers. Thinking.

Within minutes, it had produced:

• A full executive summary with a clear “Hold” investment position and the reasoning behind it

• Market analysis across geographies, distinguishing large-scale commercial agriculture from smallholder markets

• Regulatory insight — including the EU Data Act’s structural threat to platforms built on aggregating farmer data

• A risk model with likelihood and impact scoring across ten named risks

• A 30/60/90 day execution plan with named triggers and KPIs

• A capital allocation strategy with defined release conditions

This wasn’t content.

It was consultancy-grade thinking.

The kind of work that would normally take a specialist team weeks to produce and cost tens of thousands of pounds.

And here’s the critical part.

It wasn’t just impressive. It was mostly right.

It challenged assumptions. It highlighted real risks that even experienced AgriTech investors overlook — like the collapse of FBN’s direct input model, the retreat of precision agriculture platforms from Sub-Saharan Africa after connectivity failures, and the John Deere data ownership backlash.

It even pushed back on AI hype within agriculture itself — pointing out that AI crop advisory tools achieve 85–95% accuracy in controlled trials but drop to 60–70% in real field conditions. That gap is why farmers try AI tools for one season and walk away.

This isn’t AI answering questions. It’s AI making decisions.

Tools like MatrixStrike are early signals of that shift — not polished, not perfect, but directionally important.

The real disruption isn’t AI writing content.

It’s AI replacing structured thinking.

And that changes everything.

The shift from tools to intelligenc e

Most companies are still asking: “Which AI tool should we use?”

That’s the wrong question.

The real question is: “What intelligence should we own?”

Because the future of AI isn’t generic models in the cloud.

AI that thinks the way your business thinks. Martin Lucas at Gap in the Matrix calls this “Decision Physics” — deterministic AI that produces reproducible, auditable outputs rather than probabilistic guesswork. It’s the opposite of asking ChatGPT a question and hoping for the best.

And the market is moving this way faster than most people realise. Gartner predicts that by the end of 2026:

Up from less than 5% just twelve months ago. Not general-purpose AI. Specialist agents built for specific decisions and workflows.

That’s not a trend. That’s a structural shift.

Why this matters beyond AgriTech. AgriTech just makes the argument visible. It’s complex, global, and high-stakes. So when AI performs well there — producing defensible strategic intelligence in minutes — it proves something real. But the implication is much bigger. The same shift applies to:

• Marketing strategy

• Financial planning

• Operations management

• Leadership decision-making

Everywhere that structured thinking matters.

Which is everywhere.

The uncomfortable truth. Right now:

Most employees are using AI quietly

Most leaders don’t understand how to deploy it structurally

Most organisations cannot measure its value

And yet AI is already destroying entire business models — not gradually, but all at once. Within the next two years, up to 40% of enterprise applications will include AI agents. The companies that win won’t be the ones experimenting.

They’ll be the ones who have replaced entire decision-making workflows with intelligence they own.

We are massively underestimating what happens next.

A reality check

This isn’t magic. It still depends on:

• Good data

• Clear context

• Human judgement

And most businesses aren’t ready for it yet.

But that’s exactly the point.

The Fifth Industrial Revolution moment. If the Fourth Industrial Revolution was about computers in every business…

The Fifth will be about AI in every business.

• Not rented.

• Owned.

• Trained.

• Trusted.

Because you wouldn’t rent your entire IT system.

So why are you renting your intelligence?

The winners of this revolution…

The winners in AI won’t be the ones who use it the most. They’ll be the ones who:

Structure it properly

Apply it to real decisions

Build systems around it — rather than just experiments on top of it

Because this isn’t about better prompts. It’s about better thinking.

And once thinking is scalable, everything else changes.

References

University of Melbourne & KPMG (2025) – Trust, attitudes and use of AI https://kpmg.com/xx/en/media/press-releases/2025/04/trust-of-ai-remains-a-critical-challenge.html

WalkMe (2025) – AI in the Workplace Survey (shadow AI findings) https://finance.yahoo.com/news/employees-left-behind-workplace-ai-125900282.html

PwC 29th Global CEO Survey (January 2026) – AI ROI findings https://www.pwc.com/gx/en/news-room/press-releases/2026/pwc-2026-global-ceo-survey.html

Chegg investor relations / SEC filings (2025–2026 results) https://investor.chegg.com

Gartner (August 2025) – AI agents forecast https://www.gartner.com/en/newsroom/press-releases/2025-08-26-gartner-predicts-40-percent-of-enterprise-apps...

MatrixStrike / Gap in the Matrix

https://gapinthematrix.com

About The Author

Dan Sodergren is a keynote speaker on AI and the future of work, he is an AI trainer, and author of The Fifth Industrial Revolution. He speaks to HR directors, Chief People Officers, and senior leadership teams at mid-to-large organisations across the UK.