Human-Robot Interaction Key to Cobalt Robotics Success

Cobalt Robotics' robots combine mobility with telepresence with human security specialists. Credit: Cobalt Robotics

August 23, 2018      

In the physical security market, robots have been added to help companies supplement their human security guards with mobile platforms that can autonomously monitor a location with additional sensors to trigger alerts back to a central location. But many of these robots have lost the human interaction needed to react quickly to events and analyze situations as they occur. One company blending mobile robots with human operators is Cobalt Robotics.

Travis Deyle, Cobalt Robotics

The San Mateo, Calif.-based company secured $13 million in Series A funding in March 2018, led by Sequoia Capital. Combining a mobile robot with a telepresence-like interface, the company’s systems allow human “Cobalt Specialists” to solve problems that robots cannot.

Robotics Business Review recently spoke with Dr. Travis Deyle, the company’s co-founder and CEO, about Cobalt Robotics, its value proposition, and its goals.

Solving the gap between robots and guards

Q: What was the motivation for the creation of Cobalt Robotics and the security robot?

Deyle: It all started with the “problem.” My co-founder and I spoke with practicing security professionals about their biggest day-to-day problems and how those problems might relate to overarching industry trends. They explained that there’s a gap in the industry between human security guards, who are friendly and responsive, but expensive, and static infrastructure like cameras, which can provide consistent awareness but can’t react to situations as they occur. They were itching for a solution to fill this gap — to provide cost-effective awareness with real-time responsiveness.

We realized that mobile robots, backed by human operators, were the perfect solution to this problem. We worked directly with those early security professionals to develop our service, realize the value proposition, and then expand coverage across their organization in a consistent and cost-effective way. In doing so, we built the Cobalt robot to bridge this gap and, ultimately, to enable people and robots to each do what they do best.

Q: What makes your robots unique compared with other robots that handle security functions?

Deyle: The most unique thing about Cobalt Robotics is our focus on positive human-robot interactions. Instead of only having an autonomous robot responding to anomalous events around the workplace, there’s a human in the loop. This allows a person to immediately respond and solve security problems that a robot cannot address.

In addition, we are focused on providing customers with a friendly robot. This is accomplished by our specialists, who have excellent interpersonal skills to provide superior security coverage and positive experiences. By melding man and machine, we can create a system that has the best of both worlds: the friendliness, responsiveness, and flexibility of a person with the superhuman sensing, unwavering attention, and the accountability and consistency of a machine.

Cobalt Robotics Security Specialists

Cobalt Security Specialists provide the human interaction through the company’s mobile robots. Credit: Cobalt Robotics

Q: A lot of robot companies give names to their robots, but you don’t – was that a deliberate decision?

Deyle: We want to minimize humanizing the robot itself. Sometimes our customers will name their robot, just like some people name their Roomba vacuum cleaner or their car. This can be an important part of cultural adoption and acceptance, but it’s also important to avoid overt humanoid designs to keep expectations in line with robot capabilities.

Instead, we put more emphasis on the Cobalt Specialists: the remote people managing security incidents. They bring human elements, like personality, to the service. By not naming the robot, or giving it other human attributes, we’re ensuring that our robot is the means for human-to-human communication between Specialists and people at our customer sites.

Q: What is the biggest misconception that people ask you about your robots?

Deyle: Here’s an example of a question we receive, “Why doesn’t your robot talk?” Most of our design decisions stem from a human-centered design mentality. If a robot speaks to you, your most likely expectation is that your reply will be immediately and accurately understood by the robot.

Unfortunately, natural language processing and machine-guided conversation have limitations today — something generally appreciated by anyone who currently owns an Amazon Alexa or Google Home device.

This mismatch is especially acute in a security context, where misunderstandings could have dire consequences. For example, during a security incident or emergency. Instead, our services emphasize human-to-human interactions mediated by a robot, wherein remote humans can talk with employees and respond to other situations that may arise during the robot’s patrols.

This example is generally indicative of our design mentality: Our product is a comprehensive service that includes hardware, software, and people. To build an effective security service, all three must operate in concert.

Q: What is the motivation for offering Cobalt as a service, instead of selling them outright to companies?

Deyle: Our goal is to provide the highest degree of customer service possible. We structured our business to provide our customers the best coverage and technology available. We’ve ensured the software that runs on the robots can be updated remotely so that we can quickly and easily equip security and facilities teams with new capabilities, like spill and leak detection.

By maintaining communication with our customers, we continue to learn more about their pain points in order to develop new features and solve pressing problems.

Q: What kinds of companies would benefit most from these indoor security robots?

Deyle: Cobalt Robotics’ service isn’t just limited to security; it is more of a workplace services robot, providing capabilities across security, facilities, emergency response, and employee health and safety. We designed the robot such that it could provide valuable services in all kinds of indoor spaces: offices, manufacturing sites, warehouses, museums, hospitals, critical infrastructure sites, etc.

Our robot is particularly well-designed for workspaces and human environments, so we’d like to continue honing our ability to work with people in everyday settings.

For example, the robot is covered in fabric, similar to the type you’d find in office furniture. Fabric enables the robot to seamlessly blend into its indoor space and enables us, as roboticists, to have flexibility about the kinds of sensors and hardware that we install to meet the needs of security teams.

Q: What has been the biggest challenge for Cobalt Robotics in its early stages?

Deyle: To us, it’s important to provide a service that combines the best attributes of humans and robots to solve a problem. In the early stages, the biggest challenge was meeting the demands customers have for how they’d like to design their security toolkits.

Cobalt is designed to help them fill in the gap between static infrastructures like cameras and the friendly, reliable security guards who can respond to situations. Cobalt Robotics has created a customer-facing global security operations center [GSOC] so that security teams have visibility into robot-detected anomalies and a direct line to Cobalt’s robot Specialist.

As the specialist plays many roles such as concierge and security, a big challenge is finding outgoing, friendly people who are willing to work nights and weekends.

Cobalt Robotics robots

The company provides robots as a service to companies. Credit: Cobalt Robotics

Q: Do you anticipate adding more functions to the robot – additional sensors, better cameras, etc.?

Deyle: We’re always open to customer feedback about what will help them improve their security toolkit. Cobalt has the ability to run software updates remotely so security teams have robots that are constantly updating to include new capabilities such as spill and leak detection.

Q: At the end of March, you raised $13 million in Series A funding. What are Cobalt Robotics’ plans for the funding?

Deyle: Our goal is to grow the team and scale production. We’re looking to service more customers in the geographies where we’re already located and are assessing opportunities to continue expansion. Our team is also growing across technical and nontechnical roles. We moved offices in July and are already quickly outgrowing that space.

Q: What are robot companies getting right these days? What are they getting wrong?

Deyle: Some of the most interesting robotics companies, like Fetch, Veo Robotics, and Iron Ox, are ones that examine a problem and apply robotics as a solution to solve it. The best companies we see doing this put people at the heart of the experience and prioritize human-robot interactions.