AI for Retail

CHAPTER 13A “General Theory for Retail AI”

Now that you have a solid AI foundation, we can discuss our general theory of how to apply AI to retail.

Retail stores are incredibly difficult to operate. Retailers have to select, manage, price, and keep in stock 4,000 to 100,000 SKUs in every store, predict how much to order next, and ensure the right product came in the night before. They have to ensure their staff are trained up, doing the right thing, the right way, and in the right order. They have to deal with thousands of customers coming into each store each day and make sure they get out the door quickly and with a smile on their face, and do so with less and less labor each year as the price of labor goes up and the availability of that labor goes down. Add on top of that supply chain shortages, inflation, deflation, e‐commerce, the entrant of new digitally native competitors, and quickly growing organized retail crime (ORC). This is a tough, tough business.

I personally cannot imagine running all of this even close to optimally without AI. After seven years working in retail, I have seen some really shocking stuff. Stocker processes where half of the product that shows up on trucks shouldn't have been ordered in the first place. Price tags that are not keeping up with inflation. E‐comm substitutions as high as 30%. Eighty‐five percent on‐shelf availability where over half of the outs are not orderable, and the retailer shrugs saying, “What can I do? They aren't shipping?” ...