News of the day

Our take

Hi Dotikers!

Fortune sums up a stinging finding: according to a new MIT report, 95% of generative AI pilots aren’t producing tangible impact in business. The key issue isn’t model quality but the learning gap between the tools and the organizations. Another takeaway: companies that buy off-the-shelf solutions fare better than those tinkering in-house.

This signal isn’t isolated. S&P Global notes a sharp rise in AI project abandonments: 42% of companies have scrapped most of their initiatives this year, up from 17% last year. And on strategy, MIT Sloan Management Review reports that the share of organizations that consider AI core to strategy fell back to 38% in 2024. We’re seeing methodological fatigue, not a technological breakdown.

Our take: the problem is execution. As long as AI remains a gadget run by showcase POCs, it won’t touch the P&L. If your POC fits on an animated slide, it’s probably a PowerPoint POC. Winners rewrite their workflows, set measurable goals, give business teams a real product mandate, and lean on proven solutions when that’s faster than building everything from scratch. Data is treated as an asset, governance sits at the CEO level, and deployment follows workload demand rather than the trend of the moment. McKinsey’s work points the same way: organizations that “rewire” their processes for AI see more impact, while only 1% call themselves truly mature.

So: stop the decorative experiments, target repetitive, measurable use cases, then industrialize 😉 AI won’t save a confused organization. A clear one will know where AI creates value and how to extract it.

Meme of the day