Starbucks implements a recommendatory system with artificial intelligence from Microsoft

On May 7, 2019 the Microsoft company reported that Starbucks for improvement of work of the coffee shops uses technologies of Microsoft.

Starbucks uses technology of training with a reinforcement (reinforcement learning) based on a cloud platform of Microsoft Azure for formation of the personalized recommendations of food and drinks for more than 16 million users of the mobile application of Starbucks Rewards. For example, if the visitor constantly orders drinks and food in which there is no milk, the platform will draw conclusions on his preferences and will offer it products, this ingredient is not in which.

As of May, 2019 a system is available only to users of mobile application, but in the company already test technology of recommendations which also clients who do the order of cars will be able to use. In the special digital menu for them the recommended products will be highlighted. A system will create recommendations on the basis of 400 different factors, considering, in particular, weather conditions, time of day, availability of ingredients in specific coffee shop and what dishes more often were selected by other clients. The system of recommendations in the innovation center Tryer Lab in Seattle is tested. Starbucks is going to make it available in the coffee shops in the nearest future.

Also Starbucks implements technologies of Internet of Things for ensuring smooth operation of coffee shops. In each coffee shop there is a set of different devices: from coffee machines to blenders which work approximately for 16 hours a day. Failures in operation of the equipment can significantly increase expenses on service maintenance and stop work of coffee shops. For reduction of number of breakdowns in Starbucks decided to connect devices to the operating system of Internet of Things Microsoft Azure Sphere.

The coffee machines connected to Internet of Things analyze a set of factors in preparation time of each cup of espresso: from a sort of coffee beans to structure and water temperature. For eight-hour change more than 5 megabytes of information collect. To connect different devices of Starbucks to Azure Sphere and it is safe to store data, Starbucks uses the protective module (guardian module) of Microsoft. It will allow the company to use model of predictive service of machines: to beforehand reveal problems which can arise in operation of the equipment and predict time in which it will be necessary to order service maintenance.

The device will also reduce delivery time of new recipes in coffee shops. As of May, 2019 specialists carry instructions for preparation of new drinks on points and manually install software in each coffee machine from the flash drive. Thanks to system implementation, recipes will be loaded into coffee machines from a cloud on a one click of the button.

Also Starbucks works on one more option of mobile application which will allow clients to obtain information on coffee from the pack purchased by them: where it was grown up as the company supports farmers in this region where and when grains were fried also many other things. Such transparency will be provided by Microsoft Azure Blockchain Service. Read more here.