Get out of the doing

Many accounting teams are so buried in booking entries that nobody has space to interpret what those numbers even mean. Everyone processes their piece correctly, but nobody looks at the whole. The problem only surfaces when executives review the numbers. The executive level should never be the first team noticing a loss ratio reaching 90; the expert accountants should flag it explicitly to them with guidance on what to do about it.

Large companies split work into layers. One person does an entry, one reviews it, one thinks about what it means. You can continue nesting work like this as companies grow, and it means a very small percentage of the overall team actually interpret the numbers.

Smaller teams are hit the hardest. They aren’t big enough to split work like this, so when they are barely meeting the deadline, they have no time to use their domain expertise and interpret why the results are as they are.

With AI, we can change this. One person can own the full slice: run a report, review it, interpret it. In just a few months we’ve saved 30 hours a month, which is dramatic for our 3-person team. What’s more important is that we now have the space to use our actual expertise and give leadership guidance on what needs to happen. This finally lets us spend our time on strategy like we were actually hired to do.

I’ve found that the only way to effectively use AI in this kind of work is when you understand the fundamentals of the operation well enough to notice when it goes wrong, and give it instruction to improve it for next time. If you don’t have this experience - either as an engineer who hasn’t done accounting work, or as a middle-manager who hasn’t had their hands in the actual work - you’ll have a much harder time.

The “doing” was never supposed to be the focus of our job. It was just a necessary step along the way. I’m excited for us to move on to more thinking.