AI PILOTS
Only 5% of AI pilots are making money, yikes
A new MIT report finds that while generative AI holds promise, most enterprise projects aren’t paying off.
Only 5% of pilots deliver rapid revenue growth.
The rest stall, despite heavy investment.
Startups are the outliers - some have gone from zero to $20 million in just a year by fixing one pain point with AI and partnering smartly.
For larger firms, the issue isn’t model quality but a “learning gap.”
Tools don’t adapt well to workflows, and organisations struggle to integrate them.
Budgets are also misaligned: most spending goes into sales and marketing, but the best returns come from automating back-office work like outsourcing and admin.
Three things to note from MIT’s research:
Only 5% of pilots boost revenue.
Startups thrive by narrowing focus.
Bought-in tools outperform internal builds.
Flop era stats
Buying from specialist vendors works around two-thirds of the time; in-house builds succeed only a third as often, though many firms still try.
Success also depends on giving managers, not just central labs, the power to drive adoption.
Workforce shifts are already visible in admin and support roles, often through attrition rather than layoffs.
Meanwhile, “shadow AI” use continues to spread.
Looking forward, advanced firms are testing “agentic” AI systems that can learn, remember, and act within limits.
MIT basically said: stop DIY-ing AI, you’re bad at it.