September 28, 2022
By Keith Shaw
Founded in 2016, Cobalt Robotics develops mobile security robots that work alongside human security professionals to monitor indoor corporate office locations with services such as patrolling, leak and spill detection, alert monitoring, and other security guard tasks. A unique feature of their robots is a two-way video interaction system that connects the robots to human security officials who can speak with employees or other people to react to potential security incidents.
The company recently announced partnerships with other companies to address mobile robot integration in buildings – it works with Otis Elevator to allow robots to access elevators remotely, and Safetrust to integrate its security sensors into the mobile robot platform. The company also recently launched an enterprise video platform option that connects staff and visitors with a live employee from the customer’s organization or security team first, with backup from Cobalt’s Command Center.
Robotics-World recently spoke with Mike LeBlanc, Cobalt’s Chief Operations Officer, about the growth of the company and the importance of partnerships with other companies in the security and office building management space.
Robotics-World: Talk about some of the reasoning behind recent partnerships that you made – did you approach the companies or did they come to you?
LeBlanc: We’ve been working on this for a number of years to see how we fit within the security ecosystem. What we realized more and more by listening to our customers is that they do not want a robot vacuum-like appliance for their security space – they don’t want something that has its own dashboard, or have to look at a new feed and pull up new systems in their secure operations centers. They want something that already ties in with their existing systems. So this is something that we’ve worked on in different ways, with open APIs that different companies offer for years. But now we’ve realized the power of engaging at higher levels with the companies themselves, and making it a joint project, to figure out how we can get the most out of their product and the most out of ours.
Safetrust is a great example – they do identity and all varieties of access control. They’re replacing a lot of legacy systems for figuring out who’s in your building and whether they should be there. Our robot can have a mobile reader on it that ties in with Safetrust, to not only catch people when they’re going in and out of an access point – like a door – but actually just when they’re working in the space. So you can go outside your IDF [Intermediate Distribution Frame] room and see whether that person actually belongs. Is there someone milling about or loitering that shouldn’t be there? Is it somebody who is in a Zone 4 space that is only cleared for Zone 3 – it’s those things that the robot can badge challenge, and then actually have a person come and interact with the other person to find out who they are and if they should be there. So that’s a very powerful thing – it’s not only good for us, but it’s very good for Safetrust as well.
Otis is another big partnership – if you put us in a building, we go mostly in Class A high-rises, and we can only cover one floor. It’s going to be extremely expensive to have us cover the same space that a guard would. So being able to go up and down with robots – it’s a core competency of ours. We’ve done this in a number of different ways for the past several years, but now by forming partnerships like the one we have with Otis, you really start to see that we can get past some of the technological issues to really be able to say these robots can scale, they can go anywhere. There’s an Otis elevator – we can get the robot to move up and down between floors. That’s a very powerful message to clients.
R-W: Is this something that you realized that in terms of getting from one floor to another, you would have had to develop a manipulator arm or other button-push device on the robot?
LeBlanc: If you came to our offices in 2019, you would have seen as soon as you walked in a whole fake elevator with buttons and a panel, with a robot going around with a little arm designed with machine learning to go and figure out which floor to go to. The robot pokes the button – and we still have those robots deployed in some spaces, because if people have a manufacturer that we don’t work with, that’s exactly how we get to multiple floors.
But much more powerful than that is being able to call and dispatch robots through Bluetooth or other technological methods, so remotely a robot can walk up to an elevator lobby, not have to touch anything physical in the space, call the elevator, go into it and then move to the proper floor. That has been a very technological problem, but one that we have put a lot of resources toward to ensure we can do it expertly.
R-W: What does Otis get out of this partnership?
LeBlanc: They are extremely progressive in their thinking about technology. I’ve always had a bias of thinking about elevators as the stickiest product in the world. Once they’re in a building, you can’t remove them. You can’t tell them that you’re getting rid of them. What I’ve found is exactly the opposite attitude of what I expected. Otis very much wants to be part of the new workplace, they are determined to make sure they are still relevant, that people are still building buildings and realizing Otis is the way to go. So they were excited to partner with us because they can start saying ‘We can shuttle your robots around as you start to get – not just security robots, but safety facilities, cleaning robots, whatever they may be. All of the integrations we’re building are going to be able to be used for robots coming in the future.
R-W: Are you seeing, or planning for the eventuality where you will have multiple robots in a building that need to communicate – such as a security robot and a cleaning robot?
LeBlanc: Right now we’re still not seeing those situations with our customers. We target corporate security and are looking at indoor corporate offices, whereas a lot of other robotics companies are taking on challenging problems outside and in more dynamic environments. We ended up landing in corporate security because it’s such a controlled environment, and has exactly the specifications we need to build around. We realized it was a place where technology already exists to give customers a great use case. But I suspect that lots of robot companies are going to start turning to the indoor corporate office space as time goes on.
R-W: How did the pandemic affect Cobalt Robotics? If people were no longer in office spaces, did they need to have as many security guards or security tasks?
LeBlanc: The pandemic, of course, has been horrible in many ways, but it has been very beneficial to robotics automation. What we saw was that customers – rather than thinking of the 50-year-old guard model of ‘I have a building, let’s put a person in it’ – they had to rethink the fundamental question of why do they have a person there? Whenever people think about fundamental questions like that, it will be helpful for robots, because I think people had gotten comfortable in the old ways of just going along with the way the industry ran – you put guards in and that’s just part of your normal security program, and automation seemed either too hard or too expensive for them to implement. But the pandemic told people that they would have to do that hard work, no matter which way you did it. They realized it ended up being a big cost savings for them to do this.
We ended up seeing a market that said, ‘I need a duty of care to my people, I need people to feel safe, and I need to collect data.’ And they were looking at whoever could answer that question best. We started making the argument for how our robots could respond that way. That’s why we ended up seeing higher growth than we’ve ever seen through the pandemic. We’ve also been able to respond to hybrid work, remote work, flexible desk policies – such that robots can be a better solution than guards in most spaces.
R-W: So what happens when more employees return to the office?
LeBlanc: If you look at a shared service like cleaning, if you have 100 people in the building, you will likely pay 10 times more than you would if you had 10 people in the building, because the contracts usually run with how many occupants there are. For security, your duty of care does not change whether you have one person in the building or 100. In a lot of places where lower occupancy has occurred across buildings, property managers have reached out to us because they realized they can’t pay as much for security. That’s a cost they have to spread out among all the square feet of the building and the tenants they’re going to bring in. Yet they still need to see, are there leaks and spills? Are there hazardous items? Is somebody here that shouldn’t be, is there a suspicious person that we need to call the police? All of those basic questions still need to be answered, so I think they have seen that robotics can be a real answer to those questions.
R-W: What are some of the core functions that customers are looking for in a security robot? Is it having sensors and having robots do patrols, or other features?
LeBlanc: Patrols still remain core to our service, just like they do for guards. But we’re starting to see a lot of success with things like alarm response. This is exactly the place where integrations come to matter a lot, because people have their buildings decked out with sensors that will tell you if a door is held open or forced open, but the companies receiving those alarms sometimes report them in the millions across the year. Most of them are nothing. So now, the way those models work is that people dispatch a guard to go check out each one of those, because you never know which one is going to be the needle in the haystack. Robots can be automated to hear that alarm, get sent exactly to that place, and then use machine learning to determine whether to escalate so that a human needs to get involved, or just be cleared out. People are looking at this and saying there is a better way to run their programs so they don’t end up with a lot of humans in a big game of Telephone.
R-W: Why is the human-in-the-loop approach so important for Cobalt, as opposed to developing a robot that doesn’t have this interaction, where it could just travel to a location and provide a camera feed to a security professional somewhere else?
LeBlanc: This is one of the biggest differentiators for Cobalt – how friendly the robot is to employees. The interactions are hugely important. We use those interactions for things like badge challenges, asking them what they’re doing in the space, etc. This can be very helpful to the core function of security, but even more powerful is that over half the people who see a Cobalt robot go and interact with our screen without being prompted to in any way. It’s so intuitive for them to go to the face of the robot and press it, and then have someone come up on the screen who is friendly and engaged in their work. This may not be the typical stereotype of what they expect to see with a security guard. That’s extremely powerful in winning people over, and eventually gives us much better robot adoption across all our sites.
For more details on Cobalt Robotics and its technology, visit its website here.