Garbage in, Garbage out.
AI won’t solve your problems, at least not yet.
I just wrapped a project with a client whose product team is cutting-edge—but their data environment felt a couple decades behind. They were growing, yet not seeing the financial lift they should. Pricing lacked visibility. Systems were disconnected. Leaders were reading raw PDF reports with no daily signal. Classic “garbage in, garbage out.”
Step one for me is always the same: understand the business like an insider. What do they sell? Why does it win? Then zoom out: short-term and long-term goals. In high-growth companies, goals often lag behind the pace of the work. This one needed clarity.
Next, I audited the data reality: where are the leaks—in the data, in the revenue, or both? Reports were static PDFs, not decisions. So we mapped an end-to-end pipeline: pull emailed PDFs → parse clean text → stage the lines → load fact tables → publish BI that leadership can use every day.
Here’s the part most people miss. Yes, I used AI—specifically ChatGPT—to speed up the build. But AI wasn’t the hero; experience was. I’ve led Data Science, BI, Analytics, and Data Engineering teams. In some areas I’m deep; in others I know what’s possible and which tools fit. ChatGPT acted like a capable teammate: checking code, proposing fixes, helping me iterate. I still had to ask the right questions, know what “right” looked like, recognize when we were stuck in a loop, and troubleshoot hard. Without that, AI is just a fancy autocomplete.
What I learned (or re-learned):
If you don’t ask the right questions, you don’t get the right result.
If you can’t spot “this is wrong,” you’ll scale the wrong thing.
Troubleshooting is most of the job.
New tools are great—but they don’t replace judgment.
The org implications are real. With AI, managers are the new front line—orchestrating the work with a smaller analyst base. As connectivity improves, directors become the front line, then VPs. The roles won’t disappear; they’ll evolve. Companies will do more, faster, for less—if they pair AI with experience and clean inputs.
In this project, we went from noisy PDFs to a live pipeline and daily BI. Same products, same team—better visibility, better decisions, better results. That’s the payoff when you combine hard-won experience with the right questions and the right tools.
