‘Dirty Data’ Is Hurting Your Company’s AI Strategy — Here’s How This Tech Founder Is Fixing It

Most companies are rushing to adopt AI without realizing their data is too messy for machines to understand — a problem Zac Choi saw coming years ago. While others chased flashy AI applications, he focused on the unglamorous truth: garbage in, garbage out. His new venture, Big Context & Company, isn't building another AI model — it's preparing the ground so AI can actually work.

Choi makes a sharp distinction that more founders need to hear: the primary user of enterprise data is no longer human, but AI agents. Yet most data systems were built for predictable human queries, not probabilistic machine reasoning. That mismatch is why so many AI initiatives fail — not because of the AI itself, but because the data underneath is unreadable to machines. His metaphor of 'tilling the soil' before planting AI is spot-on. Too many entrepreneurs skip this step, expecting instant results from contaminated data.

What stands out is Choi’s deep operational fluency — shaped at McKinsey, Wharton, and through building and selling a previous AI startup. He didn’t just spot a gap; he’s spent two decades accumulating the rare combo of technical depth and execution experience needed to fix it. His advice to first-time founders — prioritize team alignment and distribution over product perfection — hits hard, especially his insight that you must treat the market like a co-founder.

His Zero-One-Two-Three problem-solving framework is also worth stealing: clear your mind of assumptions, break problems down to first principles, seek trusted second opinions (not just AI), and develop a 'third eye' for directionally correct decisions. In a world obsessed with speed, that balance of rigor and agility is what separates real operators from pretenders.

For any entrepreneur betting on AI, this isn’t just another founder story — it’s a warning and a roadmap. The real competitive advantage isn’t in the model you use, but in the quality of the data it runs on. And right now, most companies are building on quicksand.

Curious how to make your data AI-ready? The full interview breaks down the strategy behind the scenes.