AI in Retail and E-commerce
Last updated · 1 min read
The invisible AI you already shop with
Recommendations. "Customers also bought" is the oldest and most commercially proven AI in retail — pattern-matching over millions of carts. It is not generative AI; it is classic machine learning doing one job extremely well.
Demand forecasting. Stores predict what sells where, next week, cutting both empty shelves and waste. Grocery chains credit forecasting models for double-digit percentage cuts in food waste.
Search that understands intent. Type "warm coat for cycling to work" and the store shows fleeces and shells, not literal coats with "warm" in the name — embeddings match meaning, not keywords.
The generative wave
Product descriptions, ad variants and shopping chatbots are the new layer: thousands of listings drafted by an LLM in brand voice, chat assistants that answer "will this fit me?" from the returns database. The risk is the same as everywhere — unsupervised generated copy hallucinates specs and prices, and a wrong price in writing can be legally binding in some jurisdictions. Human review stays in the loop for anything with a number in it.