Can we use data about people to alter physical reality, even in real time, and improve their performance at work or in life? That is the question being asked by a developing field called augmented social reality.

Here’s a simple example. A few years ago, with Sandy Pentland’s human dynamics research group at MIT’s Media Lab, I created what I termed an “augmented cubicle.” It had two desks separated by a wall of plexiglass with an actuator-controlled window blind in the middle. Depending on whether we wanted different people to be talking to each other, the blinds would change position at night every few days or weeks.

The augmented cubicle was an experiment in how to influence the social dynamics of a workplace. If a company wanted engineers to talk more with designers, for example, it wouldn’t set up new reporting relationships or schedule endless meetings. Instead, the blinds in the cubicles between the groups would go down. Now as engineers passed the designers it would be easier to have a quick chat about last night’s game or a project they were working on.

Human social interaction is rapidly becoming more measurable at a large scale, thanks to always-on sensors like cell phones. The next challenge is to use what we learn from this behavioral data to influence or enhance how people work with each other. The Media Lab spinoff company I run uses ID badges packed with sensors to measure employees’ movements, their tone of voice, where they are in an office, and whom they are talking to. We use data we collect in offices to advise companies on how to change their organizations, often through actual physical changes to the work environment. For instance, after we found that people who ate in larger lunch groups were more productive, Google and other technology companies that depend on serendipitous interaction to spur innovation installed larger cafeteria tables.

In the future, some of these changes could be made in real time. At the Media Lab, Pentland’s group has shown how tone of voice, fluctuation in speaking volume, and speed of speech can predict things like how persuasive a person will be in, say, pitching a startup idea to a venture capitalist. As part of that work, we showed that it’s possible to digitally alter your voice so that you sound more interested and more engaged, making you more persuasive.

Another way we can imagine using behavioral data to augment social reality is a system that suggests who should meet whom in an organization. Traditionally that’s an ad hoc process that occurs during meetings or with the help of mentors. But we might be able to draw on sensor and digital communication data to compare actual communication patterns in the workplace with an organizational ideal, then prompt people to make introductions to bridge the gaps. This isn’t the LinkedIn model, where people ask to connect to you, but one where an analytical engine would determine which of your colleagues or friends to introduce to someone else. Such a system could be used to stitch together entire organizations.

Unlike augmented reality, which layers information on top of video or your field of view to provide extra information about the world, augmented social reality is about systems that change reality to meet the social needs of a group.

For instance, what if office coffee machines moved around according to the social context? When a coffee-pouring robot appeared as a gag in TV commercial two years ago, I thought seriously about the uses of a coffee machine with wheels. By positioning the coffee robot in between two groups, for example, we could increase the likelihood that certain coworkers would bump into each other. Once we detected—using smart badges or some other sensor—that the right conversations were occurring between the right people, the robot could move on to another location. Vending machines, bowls of snacks—all could migrate their way around the office on the basis of social data. One demonstration of these ideas came from a team at Plymouth University in the United Kingdom. In their “Slothbots” project, slow-moving robotic walls subtly change their position over time to alter the flow of people in a public space, constantly tuning their movement in response to people’s behavior.

The large amount of behavioral data that we can collect by digital means is starting to converge with technologies for shaping the world in response. Will we notify people when their environment is being subtly transformed? Is it even ethical to use data-driven techniques to persuade and influence people this way? These questions remain unanswered as technology leads us toward this augmented world.

Ben Waber is cofounder and CEO of Sociometric Solutions and the author of People Analytics: How Social Sensing Technology Will Transform Business, published by FT Press.