The introduction of robotics and machine learning is having a profound impact on the security industry, bringing together the best qualities of people and machines to provide greater security and functionality in a market that demands both innovation and evolution.
Once considered science fiction, robots are actively being deployed to flag anomalies like people in an office after hours, open doors, suspicious items, unusual sounds, smoke, and other environmental risks. Soon, security robots will be considered another tool in the security professional’s tool belt, much like surveillance cameras, access control systems and intrusion detection systems, all designed to protect people, places and assets.
While there are several providers of security robots, one recent entrant into the market is looking to change the way the security industry, and people, in general, view robots. Based in San Mateo, Calif., Cobalt Robotics has developed an indoor security robot designed to be a natural extension of a company’s security team. We recently caught up with Dr. Travis Deyle, Co-founder and CEO of Cobalt Robotics, to discuss how the company is different than its competitors and how they intend to change the perception of security robots.
Q: When you created Cobalt Robotics, what was your vision and why did you decide to create a security robot?
Deyle: While pursuing my Ph.D. over ten years ago, I designed and developed healthcare robots with arms to help people with daily living activities, such as helping older adults retrieve medication or help paraplegics shave their face. In 2016, I teamed up with a good friend from SpaceX, an aerospace company, and we set out to look for interesting problems to solve. We started to ask people in varying industries, "If we could solve a problem, what would it be?"
One of the people we spoke to was a chief security officer who said he had guards during the day but at night he could not employ any because it was too expensive. However, he really wanted to provide the high-touch, customer service interaction all the time including holidays, weekends and nights. He asked if we could develop a robot to do this. When we suggested that he put cameras everywhere, his answer was that those technologies are great, but once you are inside, there isn’t a good solution. For example, surveillance in London, where cameras are prevalent, doesn’t prevent all crime and it doesn’t necessarily provide real-time response.
We realized that robots have improved over the last decade and wondered how we could build a system to have one guard be in multiple places at once. We realized that a potential solution is to create a robot that moves around a space, builds models of what is normal and looks for anomalies, such as leaks and spills, which can then be routed to a person who resolves the incident in real time. In essence, by incorporating an autonomous car or vehicle with video, you can respond as an event occurs.
Q: What is the most significant differentiator of Cobalt’s offering over other robotic companies marketing to security professionals?
Deyle: Fundamentally, we focus on enabling human interaction. Our goal is not to just be an autonomous machine. Security events are outliers and quick response and creative problem solving are big parts of it.
Also, my co-founder and I come from a robotics background. We’ve built robots that interact with and around human beings. Those cultural aspects aren’t to be underestimated. Designing a system that can move around safely, that is friendly and isn’t intimidating and scary is a critical part of what we do. We had several criteria to achieve. Among the first was to make sure we could design a system that people would like.
Security is an interesting application, often incorporating all kinds of sensors, such as day/night cameras, thermal cameras, all the way to the badge reader. By designing a machine that is capable of integrating with multiple systems, many things can happen downstream. Cobalt’s robots use more than 60 sensors and all the sensors are there for a reason. For example, the badge reader enables us to identify an individual and determine if the person was authorized to be there.
Q: What is Cobalt Robotics go-to-market strategy?
Deyle: To date, we have predominantly worked directly with the end user. We built our first prototype and had it running after about three weeks. We wanted to work closely with the end users to make sure we are solving their problems. We are now actively pursuing relations with guard companies and integrators as channel partners. One of our recent partnerships is with Allied Universal.
Q: What do you see as some of the biggest challenges relating to robots in the security space?
Deyle: There are technology, cultural and use case challenges that need to be addressed. Robots are elaborate pieces of equipment with hardware, machine learning, artificial intelligence, and web components that enable one person to control many robots. Overall, it is a complex system. Robotics have cultural challenges due to previously deployed systems that did not perform as promised. Thankfully, we avoid these issues through our on-boarding process in which we make sure our customer’s employees are comfortable around the robot and understand its function.
Q: How are robots being used by security professionals today and what benefit do they provide?
Deyle: Our clients are in the tech, finance, defense, and manufacturing markets. Two of our most notable clients include Yelp and Credit Karma, which are currently using our robots in their Class A office space. Today, there is a gap between traditional security solutions. On one end, there is static infrastructure such as cameras and access control, which do not allow an immediate response but are cost-effective. On the other hand, there are security guards, who provide immediate response but can be costly for 24/7/365 coverage.
Customers are using robots to identify unauthorized guests. Our robots have also witnessed a slip and fall of an employee. In that instance, our Specialist was able to ask the person whether they needed help or medical attention. In another instance, an individual was breaking and entering on a customer premise. The individual didn’t have a badge so the robot escorted them out of the building and this event could be monitored by the Specialist the entire time. When a fire alarm goes off the robot can serve on the front line for an emergency response. We work with each client on how best to respond and escalate different issues.
Q: What will the security team look like in five to 10 years?
Deyle: Within the next five to 10 years, security robots will be considered just another tool in the security professional’s tool belt. While we came into this market as outsiders, we found that when security practitioners are implementing a new security program from scratch, they often employ the tried and true -- video, access control, and manned guards. These are all great systems and solutions but the costs can add up. Many new enterprise technologies have an average adoption timeline of two years. As the use of security robots grows, we need to be honest and realistic about the capabilities, but we also recognize that we have the technical expertise to deliver immediate value and long-term vision.
