ML algorithms identify "suspicious" probably overpriced items among SKUs with poor sales and manage a smooth step-by-step price reduction for ones until conversion rates improve.
As a result:
1. The item's price lowers slightly, so sales of the "suspicious" SKU raise significantly, improving the company's inventory turnover.
2. Or the "suspicious" item has no potential for good sales, even at the minimum margin level, because of poor quality, unattractiveness, bad reviews.
The platform marks it as illiquid and removes it automatically from the assortment matrix with a ban on purchase. That prevents the accumulation of illiquid stocks.