At a small cocktail stand in the middle of London’s Barbican Centre, a crowd is watching a bartender spin and whirl around a shaker full of ingredients for a passion fruit martini. The mixologist’s moves aren’t quite as acrobatic as those of Tom Cruise in , but this level of behind-the-bar showmanship is the sort of thing that would normally fill a tip jar. The thing is, though, tipping isn’t an option here. After all, what would a robot bartender spend the money on?

Makr Shakr, a robotic bartending system serving customers at the Barbican’s “AI: More Than Human” exhibition this summer, is the brainchild of Italian architect Carlo Ratti. While it started as a one-off project for a Google event in 2013, the technology is now a commercial product sold through Torino, Italy-based Makr Shakr Srl. The company has delivered robo-bartending systems to Royal Caribbean Cruises Ltd. as well as hotels and exhibitions around the world. Emanuele Rossetti, chief executive officer of Makr Shakr, spoke with Bloomberg Markets special reports editor Siobhan Wagner about how he sees artificial intelligence and data analysis playing a future role in bars and restaurants.

How does the Makr Shakr differ from robots used for more industrial purposes like car manufacturing?

The machine uses different pieces. One piece is the robot arm, which comes from car manufacturing. The arms are put together with a lot of components that we prototyped, patented, and built, such as the lemon cutter, the sugar dispenser, the shaker, and the mint dispenser.

What makes designing a robot to pour drinks rather than bolt doors onto a car more or less challenging?

If you need to produce a million cars all the same, you need less artificial intelligence. You just repeat a million times the same movement. Our machine is different because it interacts with a human being. A human being, through the app [that they download on their phone or use at a stand], decides the recipe. The person decides if he wants a lemon slice or sugar, or to shake or stir. The sequence and the movement of the robot are different all the time. There’s another difference to what the car industry does—they want to optimize the speed. Our scope was to make something that is a bit humanoid. We decided not to optimize the speed or the pure efficiency but to optimize the elegance of the movements.

What’s your manufacturing capacity?

Two years ago we made a big investment in engineering. From last Christmas we are on the market with a machine that is dramatically cheaper [currently €99,000 ($111,000), compared with the original €1 million prototype]. We are setting up a production site that will, by autumn next year, be able to produce between 70 and 80 robots per year.

Do you plan to release newer versions?

First « 1 2 » Next