# AI as leverage, not a shortcut

There's a version of using AI at work that's really just outsourcing thinking — ask the model, paste the answer, move on. I've tried to build a different habit, and it's changed how I operate enough that I think it's worth writing down.

The way I actually use it: as mental leverage. Not to skip the thinking, but to do more of it, faster, and from more angles than I'd manage solo. Test an assumption I'd otherwise let ride unchallenged. Sketch three versions of a workflow before committing to one. Push a half-formed argument until it either holds up or falls apart — on my own time, before it ever reaches a room full of people whose time costs more than mine.

At JSW One Finance, this isn't a side habit — it's load-bearing. AI is one of the product groups I lead, and some of what we're building in commercial lending right now is genuinely industry-first because AI sits at the core of it, not bolted on after the fact.

The distinction I keep coming back to: a shortcut gets you to an answer faster. Leverage gets you to a *better* answer, because you could afford to test more of them. Same tool, completely different relationship to the work.

More on the specific lending products this shows up in, once I can talk about them properly.