Your Data Isn't AI-Ready — And That's Your Biggest Lever
McKinsey reports: 88% of companies use AI. At the same time, Gartner says: 57% of organisations have data that isn't AI-ready. That's not a contradiction — that's the norm.
What AI-ready means for data sounds unglamorous: customer data that doesn't exist in three systems simultaneously with three different versions. Process documentation that's current, not from 2019. Knowledge that's captured somewhere and doesn't only live in the heads of individual employees. Permissions that are cleaned up before an AI assistant starts searching through them.
But before you clean up, you need to answer one question: which data is actually relevant? In the Handelsblatt Disrupt Podcast, Claudia Pohlink, CDO at FIEGE, asks precisely this question — and arrives at a sobering finding: most companies sit on mountains of data, the majority of which is operationally irrelevant. Only once it's clear which data actually improves decisions does the AI investment make sense at all.
In many conversations, I hear the same reflex: "We need AI first, then the data will sort itself out." The opposite is true. AI doesn't turn bad data into good results — it turns bad data into convincingly sounding bad results. And that's more dangerous than no results at all.
No company has solved this perfectly — not even the 6% that McKinsey calls "AI High Performers". The difference: these companies invest in the fundamentals before they scale AI. They redesign their workflows three times more often than others. Not because they have more budget, but because they've understood what needs to come first.
The investment that will make the biggest difference in 2026 isn't an AI investment. It's the work on your own data foundation — clarifying which data improves decisions, cleaning up, consolidating, documenting. Not glamorous, but the lever that matters.
Podcast recommendation: https://www.handelsblatt.com/audio/disrupt-podcast/disrupt-wie-finden-familienunternehmen-ihre-wertvollsten-daten-claudia-pohlink/100206062.html