Being early is expensive.

Nine years ago, I took a career break to study data science. Not to become a data scientist — I believed it would reshape finance, and I wanted to be ready.

Honestly, many of the core problems inside finance teams haven't changed much since. Senior management thinks the technology won't move that fast. Day-to-day business consumes all the bandwidth. The teams working closest to the data — the ones who'd benefit most — still struggle with basic tools like pivot tables. We're not talking about AI readiness. We're talking about basic data literacy.

The line I didn't expect to cross

In the last few months I've watched people with no coding background build working prototypes in days, automate entire reporting workflows, and extract and structure data from hundreds of documents in multiple languages.

In February, the latest models pushed me over a line I didn't expect. I wasn't just impressed — I couldn't keep up. I tested them on real finance work, and the gap between what AI can do now and what it could do even six months ago is staggering. That's what pushed me to start sharing.

What actually prepares you

What prepared me most wasn't AI skills. It was data literacy — understanding data flows end to end, where quality breaks, and how processes actually work. If you can't trace a number to its source, neither can AI. A bad process automated is just a faster bad process.

The skill that carries over isn't prompting. It's knowing your data well enough to tell when AI is helping and when it's confidently wrong — the same instinct as reviewing a junior's work.

You don't need to become a data scientist. You need to understand your data well enough to know when AI is helping — and when it's hallucinating.

And you don't need your company's permission to start.