Using IoT and AI to Enhance Workplace Safety

Reducing workplace accidents and improving safety

Buddywise is a startup with a purpose: uncovering and discovering workplace safety risks before they cause an accident. They do this with AI (Artificial Intelligence), utilizing computer vision and machine learning to improve and automate safety monitoring and risk mitigation in the workplace, most particularly in manufacturing and construction. In other words, Buddywise addresses safety risks with AI using images and video, with advanced image analytics to detect risks in the workplace.

According to ILO (International Labor Organization), roughly 2.3 million women and men succumb to work-related accidents or diseases annually, which means there are more than 6000 workplace-related deaths every single day. Additionally, there are around 340 million occupational accidents globally every year. Trips and falls can occur due to poor lighting, messy work environments, cutting corners, or being hit by a vehicle – and any of these can lead to death or serious, life-altering injuries that change the trajectory of a worker’s life and career. To put it in even more perspective, in Sweden, more people have died in workplace accidents than in the military. So, what is the number one way of reducing accidents and better ensuring the safety and well-being of workers? Prevention.

“Basically, there are two evergreen challenges in workplace safety,” explains Lamin Faye, CEO Buddywise. “The first is detecting and highlighting the safety risk, which is difficult in part because people feel they are too busy to stop and in part because you develop blind spots in your daily environment. The second one is that when a risk observation is made someone needs to take action on that, which doesn’t always happen in time. To put it in a catchier way: See something, Say something, Do something. We’re aiming to fully automate this process.”

quotes icon

Buddywise offers an end-to-end workplace safety solution that not only detects risks, but also generates alarms, reports, and feed back to workers in terms of things like lights and signals, which both improves safety and enables a response before an accident occurs.

This last bit is particularly important, because if you are identifying risks or reporting risks but there isn’t any follow up action, you’re missing an opportunity to improve workplace safety and are instead putting your employees at continued risk.

While not all workplace incidents can be prevented, most of them are preventable. To put it another way, in many cases, they could have been prevented by following practices and policies that are already in place but weren’t followed. In other cases, companies simply do not have full visibility of the safety risks in their operations and lack safe policies and procedures as a result.

Computer vision: Detection & prevention

The key to tackling workplace accidents is detecting things early. Computer vision is probably the most mature application of AI. With up to 99% accuracy rate, it is more accurate than human vision, which is an important threshold. Human vision starts off at about 95% accurate and goes down to below 90% or even 80% pretty quickly if you’re monitoring for longer than a few minutes.

The accuracy of computer vision is important because it allows Buddywise to detect things at an early stage, long before a human might notice. You also can’t rely on people to report safety incidents 100% of the time – you have to have other safety measures in place. Without that, people will continue to get injured in the workplace and the problem will never be solved.

If you look at the root causes for workplace incidents, it is easy to see patterns. Roughly 20% of machine accidents involve forklifts, so it’s a big issue. Buddywise’s AI can detect if forklifts are driving in pedestrian zones or if pedestrians are walking in forklift zones.

Lamin Faye CEO & Founder Buddywise

“It also dynamically measures the distance of people to forklifts, to see if you have a lot of close calls. Trips and slips are even more common, often caused by ‘poor housekeeping’ – people tripping over unattended objects that are left where they shouldn’t be. Buddywise detects if walking paths are blocked with objects or tools that could create a safety hazard. These are examples of some of the common risk factors we can monitor in many industries – and they’re all preventable and can be handled better and more quickly through AI in real time, allowing you to react in seconds or see patterns over time.”

Industry reaction

The reaction to Buddywise has been very positive. It’s new technology and Buddywise, which was launched in 2020, is a relatively new company. The company’s first case was implemented three months after they launched, and that company became a signed customer within six months. Buddywise’s second customer was signed shortly thereafter.

“Once we had that validation, we started building out the product,” says Lamin Faye. “Originally, we developed just a small sliver of what we actually wanted to do. We launched the built-out version at the end of last year and now we’re onboarding a number of new clients in sectors from paper and pulp to construction, which is a great response. Unfortunately, all of these companies have terrible stories that are top of mind, where if they had a safety system like Buddywise, incidents could have been avoided.”

Offering an end-to-end solution means companies don’t need to worry about the ins and outs of how the Buddywise solution works – they do what they do best and Buddywise does the rest.

“The biggest preparatory work we have with our customers is making sure we do a very thorough due diligence in terms of how we deal with customer data,” explains Lamin Faye. “Where we store it, how security works – it is extremely important to both Buddywise and our customers that we do a thorough data security assessment so that everyone is comfortable with things like privacy. This is the most important part that the customer needs to be involved in, apart from us really understanding where their challenges are. Beyond that, our solution is ready out of the box and improves continuously as the AI models learn based on site-specific data. .”

All data generated by Buddywise is encrypted, both in transit and in storage. Although they analyze data in real time, very little data is actually saved. The only time it is saved is if there is a trigger, such as someone falling down or another incident happens, which is for accident investigation purposes.

IoT & Connectivity

IoT is only as smart as the signals you give it and sensors have tended to be very single purpose. Computer vision continues to be one of the most, if not the most, critical method of connecting the real and digital worlds and is adding intelligence to IoT which makes it very useful. In light of this, Buddywise plans on integrating wearables and other devices going forward in order to further enhance worker safety.

“As IoT matures, you can see where sensors will become multi-purpose, and the possibilities to build ecosystems are enormous” says Lamin Faye. “One sensor could also be used to monitor temperature, pollution and air quality, and this is as true in a factory or warehouse as anywhere else, where building an ecosystem that uses our AI and cameras and integrates them with wearables or other types of sensors will mean an even clearer picture of what is happening in real time.”

Analyzing in real time means streaming a lot of data – Buddywise analyzes up to 30 frames per second, so that’s a lot of data. The only sensors they use currently are cameras –  Buddywise can assist in securing cameras for customers from their partner Axis, or they can integrate existing cameras. Buddywise also offers SIMs if needed – some people are happy with their existing connectivity, but Tele2 IoT connectivity is also available.

“The thing that makes Tele2 IoT stand out for us is that it has a broad company footprint in the B2B sector, and are really focused on industrial IoT, which isn’t the case with other providers,” says Lamin Faye. “Additionally, Tele2 IoT has its cloud interconnect setup that we value a lot and allows us to provide customers with secure private connectivity, which also has a very good connectivity to our servers.

“Tele2 IoT also has a global footprint and has the roaming agreements we may require going forward. Additionally, the dialogue has always been good in terms of price and other factors and Tele2 IoT has continually kept that dialogue going on a variety of issues. I’m convinced that when we move into countries where roaming agreements will be increasingly important, Tele2 IoT will have the solution we need and business models where everyone is happy.”

The future

Within the next ten to twelve months, Buddywise will have more than 1000 cameras under contract and that number is expected to double in the following year. They also will be adding more ‘bells and whistles’ to their offering not only to address the needs of their customers but to also take advantage of rapidly evolving technology.

“People often come to me with ‘wild ideas’ but as technology develops these ideas become less ‘wild’ and more reality,” says Lamin Faye. “So, as we as a company develop, we will see more ‘wild ideas’ becoming reality. A picture is ‘rich data’ and I don’t need to install a new sensor to take advantage of new technology – I just need to train the AI to detect new safety situations, so we have everything we need to evolve and that’s the really cool thing about this – we will never run out of possibilities. Some things today are really complicated, but the challenges will be solved. One current topic is detecting fall arrest harnesses, where we want to see if people are wearing it when working at height. Right now we can’t do that, but 6 months from now we will be able to do it – and that will further protect workers and increase safety in the workplace, which is what Buddywise is all about.”

Get in touch