At a recent industry event in Zürich, I had a conversation with a data analytics leader at a major global organisation. We were talking about AI, naturally — it was that kind of event. But what he said surprised me.
He wasn't excited about the latest AI tools. He was calm. Almost unbothered by all the noise around him. When I asked what was driving results at his organisation, he said: "We've been running Tableau for over ten years. Our people don't argue about the data anymore. They just use it."
That sentence stayed with me for the rest of the day.
What Ten Years of Data Discipline Actually Buys You
When a company has a mature, well-governed data environment, something subtle changes. Meetings stop starting with "but our numbers show something different." Decisions happen faster because the premises are agreed upon. Leadership trusts the reports. And when a new technology comes along — like AI — there's a clean foundation to build on.
That's the real advantage. Not Tableau specifically. Not any particular tool. It's the years of decisions that built and maintained one version of the truth.
What does that actually look like? A few things:
- One data model that all departments use — not five slightly different ones owned by five different teams
- A named owner for each dataset who is accountable for its accuracy
- A process for how new data gets defined, approved, and integrated
- Reports that people trust enough to take to the board without double-checking
None of these are glamorous. None of them are the subject of keynote talks. But companies that built this infrastructure five or ten years ago are now watching AI plug directly into it — and deliver results immediately.
The AI Trap Companies Are Falling Into Right Now
Boards are asking about AI. CEOs are committing to AI strategies in their annual reports. And so companies are buying AI tools, starting pilots, hiring consultants — before they've asked the most important question: what data will this AI actually run on?
I've seen this pattern too many times. A company implements an AI-powered forecasting tool. It pulls data from Salesforce. But the Salesforce data is inconsistent — different reps define deal stages differently, some opportunities haven't been updated in three months, the customer names don't match what's in the ERP. The AI produces a forecast. Nobody trusts it. The pilot fails. The conclusion: "AI doesn't work for us."
The AI worked fine. The data didn't.
The uncomfortable truth: AI is a multiplier. If your data is a 3 out of 10, AI gives you a better 3. It doesn't transform 3 into 8. The companies winning with AI aren't the ones who invested most in AI. They're the ones who invested most in data — and are now letting AI compound that investment.
The Sequence That Works
There's a reliable order of operations for companies that successfully get value from AI. It's not the order that gets talked about at most events, but it's the one I've seen work.
First, identify what decisions matter most in your company. Not what data you have — what decisions drive outcomes. Revenue, margin, capacity, customer retention. Start there.
Second, trace each of those decisions back to the data that drives it. What do you need to know to make this call confidently? Where does that data live? Who owns it? Is it clean?
Third, fix the data for those decisions before you touch anything else. This is the unglamorous work — de-duplicating records, aligning definitions across systems, establishing ownership, building a governance process. It takes months. It is completely worth it.
Fourth, build the reports that those decisions require. Not hundreds of dashboards. The five or ten views that leadership actually needs to run the business. Make sure people trust them. Verify them against reality until the trust is earned.
Fifth — and only fifth — start exploring what AI can add on top of that foundation.
The Best Time to Start Was Five Years Ago
I know this is frustrating to hear if your data infrastructure is not where it should be. The companies that started ten years ago have a real advantage that you can't shortcut your way past.
But there's a version of this that's still worth doing. The companies that start building their data foundation today will have a compounding advantage in three years. The ones that don't will still be explaining to their boards why the AI pilot didn't deliver.
You can't skip the foundation. You can only decide whether you start building it now or later.
The organisations I met at that event who were most confident about their future weren't the ones with the most impressive AI demos. They were the ones who, years ago, decided that reliable data was a strategic priority — and never stopped treating it as one.
That's available to any company willing to make the same decision.