The other day, on my way back home after delivering a full-day AI workshop for executives, a billboard caught my eye: “Make everyone a data person.”
Since then, I kept thinking about it because every AI conversation eventually runs into the same bottleneck: data.
Most organizations still treat data as something that belongs to IT or the data team. But today, every employee is a data person: The sales rep updating the CRM, the project manager logging status, the support agent recording a ticket, the engineer writing documentation, the operations team capturing process details…
They’re all creating raw ingredients that AI depends on.
However, the problem is that almost nobody thinks of themselves that way: I’m a data person. And as a result, almost nobody feels responsible for the quality of the data they create.
The Conversation That Never Ends
Since ChatGPT launched in 2022, I’ve heard executives in every industry say the same three things:
-
“Our data is a mess.”
-
“Our data is scattered everywhere.”
-
“We need to fix our data before we can do anything with AI.”
And then… nothing. 😐
Three years later, same conversation. The data is still a mess. Everyone is waiting for the right moment, the right budget, the right strategy to fix it all at once.
The data problem isn’t new. We’ve spent decades talking about data quality, data warehouses, governance, and interoperability. AI just put a spotlight on an issue that was already there.
The surprising part is that the spotlight hasn’t created much urgency!
Now the Cost Is Visible
The impact is becoming harder to ignore.
Organizations continue investing in AI, but many AI projects struggle because the underlying data isn’t ready. And interestingly, leaders know data is often the blocker.
So why haven’t many companies started fixing it?
Why Companies Get Stuck
When executives say “We need to fix our data,” they’re often talking about decades of scattered, undocumented, and inconsistent information.
Then, in reality, fixing it all feels overwhelming!
It becomes a multi-year effort evolving new systems, new governance processes, new budgets, and even organizational change. So it feels easier and safer to wait.
They wait for a data strategy, the next planning cycle, and enough budget. But in the meantime, every new action they take adds to the existing messy data pile.
The challenge is that when it comes to fixing data issues, many leaders are trying to boil the ocean. And when the problem feels that big, doing nothing and waiting feels like the only reasonable response.
But the good news is that AI doesn’t require you to fix everything, at least not at once. It requires you to fix what matters first.
Three Moves That Turn the Ocean Into a Glass of Water
1. Stop being paralyzed by fixing the past. Start fixing the flow.
The historical mess is largely behind you.
But imagine if, back in 2022, you’d said: “From today, everyone delivers data in this format, to this place, to this one standard.”
You’d have at least three years of cleaner data today.
Instead of treating data quality as a cleanup project, focus on where data is created. Put standards, validation rules, and ownership at the source.
A practical way to do this is through data contracts: simple agreements that define what good data looks like and who is responsible for providing it.
Most organizations are still mopping the floor. The better approach is to fix the leak.
2. Let your AI strategy guide you on what gets fixed first.
One of the biggest misconceptions is that AI requires all data to be cleaned up before anything can move forward.
It doesn’t! You only need to improve the data required for the AI use cases you actually plan to pursue.
Start with the business outcome of an opportunity. Then identify the data that the outcome depends on. Fix that first.
The unit of work shouldn’t be your entire data estate. It should be a specific use case. That’s a much smaller and more achievable problem.
3. Make everyone a data person on purpose.
This is where the billboard got my attention. The most successful organizations are moving toward federated ownership, where the teams that create data are responsible for its quality.
Marketing owns marketing data. Sales owns sales data. Operations owns operational data.
The people closest to the data are usually best positioned to improve it. More importantly, ownership becomes clear. Because when everyone is responsible, no one is responsible.
“Everyone is a data person” doesn’t mean everyone becomes a data engineer; it means every team understands that the information they create is an asset on which the organization’s future depends.
One caution: don’t try to assign ownership for all domains at once.
Start with the two to three domains most critical to revenue, operations, or your priority AI initiatives.
A Simple Strategy
You don’t need a perfect strategy to get started. You need one repeatable loop:
-
Pick one high-value AI use case.
-
Define what good data looks like for that use case.
-
Assign ownership to the team closest to where the data is created.
-
Set the standard going forward.
-
Measure results and use the success to fund the next effort.
That’s it. Small steps to gradually fix data issues.
So What?
The data will never be perfect.
If your organization is waiting for perfect data before moving on to AI, you’ll be waiting forever. The simple, practical approach to making progress is to solve the data problems that matter for the next AI use case, rather than solving all data problems first.
That’s a much smaller step, but more practical. The billboard had a point: everyone is a data person.
Pick one use case. Define one standard. Assign one owner. And then start.
Because every day you wait, you’re creating more of the data debt you’ll eventually have to pay.
❓So, what’s stopping you from starting today?
If you’re navigating this shift in your organization or just curious about what’s next, I’d love to connect. I’m always interested in exchanging ideas, learning from others, and helping leaders think through the road ahead.
🤝 And if you find this article helpful, feel free to share it with your network.