The laser pointer is vibrating. I’m watching the red dot hover over a bubble labeled ‘Hyper-Personalized AI Ecosystem’ on slide 34 of the quarterly strategy deck, and the man holding the remote is sweating through his 234-dollar shirt. There are 54 people in this room, most of them nodding with the rhythmic precision of those bobblehead dogs you see on the dashboards of taxis. It’s a beautiful slide. It has gradients. It has icons that suggest interconnectedness and seamless flow. But I can’t stop thinking about the fact that I spent my morning in the basement of this very building, where the server rack sounds like a jet engine trying to swallow a bag of gravel, and the operating system is still running on a version of Windows Server 2004 that hasn’t seen a security patch since the glaciers started melting in earnest.

I recently spent 14 hours matching every single pair of socks in my dresser-an act of obsessive order that left me feeling strangely powerful-and I realized that most data strategies fail because they are the equivalent of buying a velvet tuxedo while your actual feet are covered in blisters because you refuse to buy new socks. We want the prestige of the AI-powered future without the manual labor of the present.

Aisha D.-S. sees the world in terms of flow and obstruction. When I explained to her what a ‘data strategy’ usually looks like, she laughed until she nearly dropped her inspection camera. To her, a data strategy that focuses on ‘predictive insights’ without addressing the 444 disconnected legacy databases is like trying to install a high-efficiency wood stove into a chimney that is currently housing a family of very confused raccoons. You can’t just wish the smoke away. You have to climb onto the roof and get your hands dirty.

In 14 years of inspecting chimneys, she has never seen a fire start because the mantle wasn’t pretty enough. It’s always the hidden stuff. The soot. The cracks in the liner. The things nobody wants to pay for because nobody can see them from the living room.

We are currently obsessed with the mantle. We want the generative AI to write our emails and the machine learning to predict which customers are going to churn by next Tuesday at 4 o’clock. But our data is currently trapped in a series of Excel spreadsheets that are being passed around like illicit substances in a 1980s nightclub. We have 84 different versions of the ‘truth,’ and yet we are sitting in boardrooms talking about ‘single source of truth’ as if it’s something we can just buy off the shelf and plug in. It’s a wishlist. It’s a letter to Santa Claus written by people who should know better.

I’ve made this mistake myself. I once convinced a client to spend 64 thousand dollars on a dashboarding tool before we had even checked if the data being fed into it was accurate. The resulting visualizations were beautiful, glowing in 4 shades of neon blue, but they were essentially lying to us in high definition. It took me 24 days to admit that I had prioritized the ‘look’ of the strategy over the ‘logic’ of the execution. I felt like a fraud, and I probably was one. I was selling them the marble mantle while the chimney was still full of raccoons.

A real strategy doesn’t start with ‘What can we achieve?’ it starts with ‘What is currently broken?’ This is the pivot that most leadership teams are terrified to make. Admitting that your data infrastructure is a disaster is an admission of past neglect. It’s much easier to talk about the 4-year plan for AI dominance than it is to admit that you need a 14-month project just to clean up your customer master record. But the secret that companies like

Datamam understand is that the cleanup is where the actual value lives. You can’t build a skyscraper on a swamp, no matter how much you spend on the gold-plated faucets for the penthouse.

There is a strange beauty in the plumbing. When you finally get the data flowing correctly-when the pipes are clear and the pressure is consistent-the ‘AI’ stuff actually becomes easy. It’s almost boring. If your data is clean, accessible, and structured, the algorithms work. It’s not magic; it’s just math acting on high-quality fuel. But we treat the fuel like an afterthought. We treat the data engineers like the people who come to fix the toilets-necessary, but unglamorous. We would much rather have a ‘Chief AI Officer’ than a ‘Chief Data Quality Officer,’ even though the latter is the only one who can actually make the former successful.

I think back to Aisha D.-S. standing on my roof. She wasn’t looking at the view. She was looking at the mortar between the bricks. She told me that most people wait until they smell smoke before they call her. By then, the damage is already at a 4-alarm level. Corporate data strategy follows the same trajectory. We wait until a major audit fails, or a product launch collapses because the customer data was 54 percent inaccurate, and then we scramble. We call for an ’emergency transformation.’ We hire consultants who give us even more slides with even more bubbles.

What if we stopped? What if, instead of adding another layer of ‘insight’ on top of the rot, we spent the next 14 weeks just mapping out the rot? It wouldn’t be a popular move. The shareholders wouldn’t get a press release about it. But it would be a strategy. A strategy is a choice to do one thing at the expense of another. If your strategy includes 44 different priorities, it isn’t a strategy; it’s a grocery list for a party you can’t afford to host.

I’ve noticed that since I matched my socks, I spend 4 minutes less every morning getting ready. It’s a small, almost invisible optimization. But over the course of a year, that’s 1460 minutes, or roughly 24 hours of my life back. Data strategy is the same. It’s the small, boring optimizations-the naming conventions, the API protocols, the deduplication logic-that eventually compound into the ability to do the ‘big’ things. You don’t get to the moon by focusing on the moon; you get there by focusing on the metallurgy of the fuel tank.

We need to start rewarding the people who find the cracks in the chimney. In most companies, the person who points out that the data is garbage is treated like a buzzkill. They are the person who ruins the ‘vibe’ of the innovation meeting. But that person is your Aisha D.-S. They are the only one standing between your ‘Hyper-Personalized Ecosystem’ and a very expensive, very public fire.

If you want a strategy, stop looking at the slides. Go find the person in your organization who is currently holding a metaphorical brush and covered in digital soot. Ask them what’s blocking the flow. Don’t give them a new AI tool. Give them the authority to clear the chimney. It won’t look good on a PowerPoint, but it will keep the house from burning down, and in the end, that is the only metric that actually matters.

We have enough wishlists. We have enough dreams. What we need is more people who are willing to admit that the most important work is the work that no one will ever see until it stops being done.