Intelligent efficiency companies operating in the commercial building sector tend to fall into two broad categories. Either they install lots of sensors and meters through a hardware-intensive approach, or they rely purely on software-based analytics to automate a building. 

But very rarely do companies make building occupants a central piece of how the product functions.

Building Robotics, a startup formed by two computer scientists at the University of California, Berkeley in 2012, has built an energy management app based entirely on that human-centric premise. The company's product, called Comfy, allows individuals within a building to control the temperature immediately around them -- turning them into human sensors who feed data back to a central learning algorithm.

As Building Energy gathers more information on preferences within different zones, it can then optimize HVAC use to automatically accommodate for the variations.

Anyone who's ever worked in an office building knows how fickle humans can be with temperature preferences. Can an app that keeps people comfortable also work to save energy in a building?

Investors believe it can. Last November, Building Robotics closed a $1.1 million seed round from Claremont Creek Ventures, Google Ventures, Formation 8 and Red Swan Ventures to prove out the concept. And this week, the company announced the closure of a $5.5 million Series A round led by Claremont Creek Ventures and The Westly Group to continue expanding.

"Over the last ten years or so, we've been interested in algorithms to control buildings. We looked at both increasing sensors in buildings and running fancy algorithms to automate settings. I’ve never seen anything [like what] Comfy does," said Nat Goldhaber, a co-founder of Claremont Creek Ventures, in an interview.

When Building Robotics announced its seed round last November, it didn't release any data publicly about how the Comfy app was performing. CEO Andrew Krioukov only said that the company's performance was in line with a UC Berkeley study concluding that optimization techniques could cut energy consumption in commercial buildings by 30 percent.

Along with announcing its Series A, Building Robotics is now releasing a bit more data on performance. According to Krioukov, pilot projects at a Google office and a large federal government building have brought 22 percent reductions in energy consumption.

At Johnson Controls' headquarters in Milwaukee, Wisconsin, the app has reduced energy consumption by 23 percent. And more than three-quarters of occupants in the building are using the app. 

"This shows how collecting data from people can both empower users and improve efficiency," said Krioukov.

In one pilot, a group of users wrote bots to automate the Comfy app based on their preferences. 

"We take these kinds of signs as evidence that people's excitement for controlling their physical surroundings is growing, and we are exploring releasing a public API to allow people to write new apps for their offices," said Krioukov.

Selling new software into buildings isn't easy. Building managers are typically skeptical of new technologies that will take control away from them. And because there are so many apps, dashboards and analytics offerings for buildings on the market, breaking through the noise can be tough.

But Building Robotics thinks it has a unique selling point: rather than selling efficiency, it's selling individual comfort.

When the startup approached Johnson Controls -- a leader in building controls and efficiency services -- to integrate the system at one of its offices, Building Robotics encountered lots of skepticism. But the same people who were most critical eventually became the technology's biggest supporters, said Goldhaber.

"All of a sudden, there was a drop in calls from people complaining to the building manager about temperature," he said. "Those requests almost disappear when there's an installation."

Building Robotics will use the $5.5 million to improve the app, grow its sales team and get beyond pilots. Sometime in the next couple of months, Oak Ridge National Laboratory will also release more data on how the company's human-based approach to efficiency works.