A recommendation engine is software that suggests products, menu items or content to a customer based on what they and similar customers have viewed, bought or rated. It powers the "you may also like", "frequently bought together" and "complete the look" panels seen on most online stores and apps.
It works by finding patterns in data. Collaborative filtering looks at what people with similar histories bought. Content-based methods match product attributes such as category, size, colour or ingredients. Most modern systems blend both with machine learning, and many also weigh context such as time, device and current stock.
Retailers place recommendations on product pages, in the cart, in emails and on receipts. Restaurants and cafés use them in ordering apps and kiosks to suggest a side, drink or dessert that suits the item already chosen. Well-placed suggestions raise average transaction value without discounting.
Check what data the engine needs and whether you have enough of it: new shops with little history get weak results at first. Make sure it respects stock levels, margin and dietary or allergen rules, and avoid suggesting items the customer has just bought. Test against a plain best-seller list to confirm it genuinely performs better.
For small and mid-sized businesses, recommendations are one of the most accessible forms of AI, because the data already sits in the POS and ecommerce platform.