This is an interesting article on how one company is adapting to fluctuations in its underlying costs. But here's the part that really caught my eye:

The plan was put into motion this year after the company revamped its data collection on sales, input costs and other items. Such figures used to come with a delay, and often weren’t granular enough to make decisions about specific products. Now, the company monitors sell-through, inventory trends and product performance in real time and at a store level, and executives can see the margins tied to each product, Cass said.

Brilliant. The world is in love with the fancier side of AI these days, but nothing beats plain old data fundamentals.

(Bonus: once you've sorted out your first-level data collection and analysis, that paves the way to ML and AI.)