A Zondits Interview: Building Robotics Optimizes HVAC Controls with Comfy

Michael Dean for Zondits, August 19, 2015. Image credit: Building Robotics

Whether it’s walking in from a hot summer day to a freezing cold desk or in to an overly heated conference room in the dead of winter, employee complaints about temperature are prevalent in the workplace. But Building Robotics has developed a software for HVAC systems called Comfy that provides an easy-to-use interface for employees to control the temperature of their office space. With the controls accessible through a computer or mobile phone, user requests for heating and cooling are easy to make, and Comfy caters to all areas from open office spaces to conference rooms through a simple forced-air convection system. Zondits spoke with Anna Lui of Building Robotics to learn more about what goes on behind the scenes with Comfy.

The Comfy system must be accompanied by an HVAC system that can respond to each individual request. What are the requirements of the HVAC system to meet these user requests?

Most commercial buildings utilize some type of a building management system (BMS), a computer­-based control system that controls the building’s mechanical equipment, such as heating, cooling, ventilation, lighting, security systems, etc. Comfy connects to a building’s existing BMS remotely via BACnet, the standard industry communication protocol. Comfy works best with variable air volume (VAV) systems or other forced air systems. There are some less common systems that use radiant heat that don’t work as well because they can’t quickly respond to controls.

How can Comfy deal with an open office area with conflicting user requests?

This is oncomfy-appe of the most common questions we receive. First, we’ve found that people actually share very well when they use Comfy! We’ve found that 95% of the time people in the same zone agree on the temperature setting. Comfy gives users greater insight into their personal preferences compared to the preferences of their neighbors, we’re able to avoid possible conflict.


Secondly, one of the reasons we see people disagree a lot is because there is only one temperature setpoint for their area for the entire day. Comfy adjusts continuously throughout the day. As illustrated in the graph below, people who “run cold” can get warmer via Comfy in the morning and people who run hot can get cooler via Comfy in the afternoon.


Finally, in offices with a lot of people in one zone, we offer Buddy Mode. For example, our client AppNexus in New York has a ton of open office space. In some places, hundreds of people’s desks are in one zone. We activated Buddy Mode in the open office spaces, requiring at least two people in the same zone to make the same request before Comfy triggers the response. In conference rooms, we still only require one person to make a request for an immediate response to kick in.

Does this system implement any learning features to better access the consumers’ comfort conditions?

Yes! Comfy is the world’s first intelligent thermal control for commercial buildings. Our patented machine­-learning algorithms, first developed while our founders were completing their PhDs at UC Berkeley, learn from individual requests and dynamically adjust setpoints for optimal energy savings and comfort. Comfy is able to fine-tune the setpoints on a zone-­by­-zone basis to ensure that the HVAC system is using just the right amount of energy where and when people need it. No more. No less.

Your case study on Johnson Controls, Inc. demonstrates the success of the system in an office layout building. What other types of buildings/areas does the Comfy system excel in?

Offices are where we’ve focused our interest initially, but Comfy is applicable in other types of buildings as well. Comfy’s ability to solve over air­-conditioning and freezing workplaces during the summertime has especially resonated with a lot of employers. Because we are able to deploy Comfy on a zone-­by­zone basis, it is also commonly deployed in sections of mixed­-use buildings, like ones owned by universities and the government. As we roll out more features in the coming months, we will be exploring other building types as well.

In the Introduction of the previously stated case study, you state “Since Comfy can be installed in only a portion of a building…” What is the reason for Comfy not being able to cover the entire building?

To clarify, Comfy is installed in a lot of complete buildings, but it’s also able to be installed in smaller areas. Part of what makes Comfy so different from other building solutions is how nimble the SaaS model is. We’re often able to deploy in a building within weeks of signing a contract and with less than 6 hours of on-­site time. Many of our clients choose to pilot on one or two floors and then roll out into the rest of their building a couple months later, which was the case with JCI.

What building or spatial characteristics inhibit the installation of Comfy?

Every building is unique, but­ we haven’t yet found any spatial limitations that our product development team can’t tackle. Comfy has built-in features to support different types of spaces, like conference rooms and larger zones and we are constantly developing new features in response to client needs. In terms of building systems, we’ve found ways of integrating with everything from radiant cooling panels to pneumatic thermostats.

Can the user expect energy savings from the implementations of this software? If so, how are these savings realized?

Buildings with Comfy have found energy savings in the range of 15%–­25%. Comfy predominantly saves energy in three ways:

  1. Comfy reduces conditioning and energy consumption when rooms, such as conference rooms and private offices, are empty.
  2. Comfy reduces conditioning and energy consumption in frequently unoccupied spaces, like elevator lobbies and copy rooms.
  3. Comfy reduces overconditioning in spaces when they are occupied.


Because Comfy allows occupants to request immediate heating or cooling in their specific zone, building operators have greater freedom to set the building’s heating and cooling setpoints much wider than usual. This is significant because a recent study conducted by University of California Berkeley’s Center for the Built Environment has shown that increasing a cooling setpoint by even just 1°F over the course of a day may reduce energy use by 13%.

What are the outlying effects this system has on the building management?

The immediate effect is happier occupants. A recent study conducted by the General Services Administration that chose Comfy as part of the Green Proving Ground program found that 83% of occupants were more or much more satisfied with their workplace environment after Comfy was installed. Many of our client reports from the last few months show even higher rates of occupant usage and satisfaction.

We’ve also seen dramatic reductions in hot/cold calls to facilities managers ­­from hundreds of hot/cold calls in a month to zero.

Finally, because we are constantly monitoring input from Comfy users and data we gather from the BMS, we’re often the first to know that equipment is malfunctioning. For example, the Comfy Concierge team noticed a high frequency of “Cool My Space” requests in a particular zone in one of the buildings we work in and reached out to the users to learn more. From those conversations, we found out that the vents in that zone were accidentally delivering very hot air when they were supposed to be delivering cool air. Comfy passed this on to the building’s facilities team who discovered that there was a broken heating hot water valve in their VAV box that was stuck open. The broken valve was still under warranty and was replaced promptly. Comfy is able to take the best of occupant feedback and machine learning to make life easier for building operators.

Building Robotics is an interdisciplinary team of engineers, designers, and building industry experts changing the way people interact with buildings. They are the makers of Comfy, the world’s first machine-learning based temperature control for commercial buildings, which provides on-demand, personalized thermal comfort in workplaces.The Building Robotics vision: ensuring that buildings meet the dynamic and diverse needs of people. For more information, visit www.buildingrobotics.com