Aug 29, 2025
Aug 30, 2025

Are 95% of AI Projects Really Failing?

MIT says 95% of GenAI pilots deliver zero ROI. Hype or reality? Unpack what the report really means, where AI pilots go wrong, and how leaders can unlock real business value.

Join our Data Community

Get the latest news in your inbox

Thank you! You are now subscribed to our news.
Oops! Something went wrong while submitting the form.

Follow us

MIT just dropped their latest report on the state of AI in business.

The media went wild. “95% of GenAI pilots are generating zero return.”

Shocking, right? The kind of stat that makes you wonder if all this AI talk is just smoke and mirrors.

But here’s the thing: once you peel back the layers, the truth is far less dramatic.

Let’s Break It Down

That 95% number comes from a mix of interviews, surveys, and analysis of pilot programs. Even MIT admits the sample is limited and could be skewed. So don’t take it as gospel.

Does that mean we should ignore it? Not at all. Because behind the headline are some uncomfortable lessons for leaders trying to put AI to work.

What the MIT Report Found (That Actually Matters)

  • Budgets are going to the wrong places – Most AI spend is funneled into Marketing and Sales. It looks shiny on paper, but the biggest ROI sits in the back office, buried in the boring, manual-heavy work nobody loves.
  • Pilots flop when they don’t integrate – Too often AI gets bolted on as a separate app. Another tool to learn. Another login. If it doesn’t slot into the flow of work, adoption collapses.
  • Build vs buy is a big call – Companies trying to build everything themselves are stalling. Those who partner with specialised vendors are moving faster and seeing more wins.
  • The learning gap is real – Everyone wants adaptive AI. What they’re actually getting? Static tools that don’t evolve. Frustration builds fast.
  • Trust disappears quickly – Roll out AI at arm’s length, without training or support, and confidence evaporates. People fall back on old habits.
  • Shadow AI is quietly delivering results – The unsanctioned tools employees adopt themselves are often outperforming centralised rollouts. Smart leaders don’t fight it. They guide it.

My "Hot Take"

Don’t get distracted by the headline. The stat might be debatable, but the findings are worth paying attention to.

If you’re rolling out GenAI:

  • Start small, but integrate it into real work
  • Chase ROI where it’s boring (admin, process, compliance). That’s where the real savings hide
  • Be honest about whether to build or buy — pride gets expensive
  • Train and support your people so the tools actually stick

The media loves a bit of clickbait. But leaders need to cut through the noise, grab the insight, and use it to make better calls. That’s the real story here.

If you need help understanding how AI can help then book in for a free AI Discovery Session with our team of experts.