For technology fans, the ideal existence resembles something out of “The Jetsons”: high-tech robots to manage everything from household maintenance to butler services. Now, it would appear that a team from MIT’s Computer Science and Artificial Intelligence Lab are bringing us one step closer: They’ve developed a collaborative team of robots that can delivery just about anything, including beverages.

Replacing human employees with robots is a somewhat macabre dream, but a popular idea among large tech companies like Amazon and Google nonetheless. The issue with robotic workers, though, is unpredictability – what happens when outside agencies prevent a robot from following it’s scripted instructions?

To address that, the MIT team developed three robots that work together to overcome uncertain situations. To demonstrate their approach, two robot “waiters” work in concert with a central robot “bartender.” The waiterbots approach people for drink orders, while the bartenderbot fulfills them. Working not unlike human waiters, the waiterbots were able to determine who asked for what, whether or not the order had been fulfilled and where orders were going.

Of course, while humans may be imperfect employees, robots can be even less-perfect – sometimes is comical (and disastrous) ways. For example, the bartending robot could only serve one waiterbot at a time, among other things.

“Each robot’s sensors get less-than-perfect information about the location and status of both themselves and the things around them,” says lead author Chris Amato, a former CSAIL postdoc who is now a professor at the University of New Hampshire. “As for outcomes, a robot may drop items when trying to pick them up or take longer than expected to navigate. And, on top of that, robots often are not able to communicate with one another, either because of communication noise or because they are out of range.”

Ultimately, these kinds of robots succeed or fail depending on how well the engineers can teach them to think more like humans and less like robots. Humans, for instance, don’t move by thinking hard about each step – they’re more concerned with possible outcomes, and the walking is second nature.

“These processes have traditionally been too complex to scale to the real world,” says Karl Tuyls, a professor of computer science at the University of Liverpool. “The MIT team’s approach makes it possible to plan actions at a much higher level, which allows them to apply it to an actual multi-robot setting.”