Gartner Says 40% of AI Projects Will Be Binned by 2027

I've seen this dozens of times. A UK manufacturer gets sold on AI — smart forecasting, automated procurement, agent-powered customer service. Six months and fifty grand later, the pilot's quietly shelved and nobody wants to talk about it.

Sound familiar? You're not alone. Gartner now predicts that more than 40% of agentic AI projects will be cancelled by the end of 2027. Not paused. Not "optimised." Cancelled.

It's not the tech, it's everything around it

Here's the bit nobody wants to hear. BCG's research on AI transformation found that success breaks down as roughly 10% technology, 20% tools and processes, and 70% people and organisational design. Ten percent tech. Seventy percent people.

Most SMEs I work with spend 90% of their budget on the technology bit and roughly 0% on the people bit. Then they wonder why it didn't work.

The numbers back this up across the board:

• MIT's Project NANDA found 95% of enterprise generative AI pilots failed to deliver measurable financial impact.
• BCG's Build for the Future study of 1,250 organisations found only 5% qualified as "future-built" — actually generating substantial AI value at scale.
• McKinsey's Global AI Survey found 88% of organisations use AI somewhere, but only 5.5% attributed more than 5% of EBIT to AI.
• S&P Global reported 42% of companies scrapped most AI initiatives in 2025, up from 17% the year before.

Four different research programmes, four different methodologies, all pointing at the same thing. The technology works fine. The deployment around it is the problem.

Why AI projects actually fail

In my experience, it's almost never the model. It's one of three things:

1. The data underneath is rubbish. You can have the best AI agent in the world, but if it's reading part numbers that don't match, supplier records that are duplicated, and stock figures that haven't been reconciled, you'll get confident wrong answers at speed. Bad data amplified by AI doesn't fail quietly — it compounds.

2. Nobody redesigned the workflow. McKinsey tested 25 attributes and found workflow redesign had the single biggest effect on whether AI delivered EBIT impact. Companies that just dropped AI tools on top of existing processes saw nothing. The ones that rewired how work actually flows? They saw real returns.

3. No clear accountability. KPMG's 2026 Global AI Pulse found organisations with clearly defined accountability for AI outcomes achieved ROI at three times the rate of those without. Not "someone sponsored it" — someone owned the outcome.

What UK SMEs should actually do

If you're a UK SME thinking about AI, don't start with the AI. Do this instead:

• Get your data sorted first. Clean master data, consistent definitions, one version of the truth. If your ERP data's a mess, fix that before you even look at AI. The agents are only as good as what they're reading.
• Pick one workflow, not ten. Find one repetitive, well-documented process — invoice processing, stock reconciliation, report generation — and redesign it properly before adding AI. Prove the model works end to end.
• Name an owner. One person who's accountable for the outcome, not just the project. Not a steering committee. A human with a name and a budget.

That's it. Not glamorous. Not what the AI vendors want to hear. But it's why 95% of AI pilots fail and the 5% that succeed are quietly getting on with it.

Don't be the 40% who bin their AI project in 18 months. Be the ones who sorted the foundations first.

References & Further Reading

1. Gartner (via Harvard Business Review), Why Most AI Investments Fail (February 2026)
2. BCG, Build for the Future: AI Transformation Study
3. McKinsey & Company, The State of AI 2026
4. KPMG, Global AI Pulse Q2 2026

Sorting your data out before you invest in AI?

I help UK SMEs get their data house in order before throwing tech at the problem. ERP data, master data, reporting trust — the unglamorous stuff that makes AI actually work.

Let's have a chat

Matty Hatton is the founder of Digital Adaption, an ERP and data consultancy based on the Wirral. He has spent 15 years delivering ERP transformations for manufacturers, including leading the data migration on a £4.5m consolidation of four legacy systems onto a single Infor LN cloud instance for a 220-user group. He holds an MSc in Digital Transformation and IT Strategy from Manchester Metropolitan University and is Microsoft PL-200 certified.

LinkedIn | Get in touch

Start with a 30-minute data risk call