As Canadian industrial workplaces move from incident reports to real-time monitoring, Voxel’s computer vision technology helps safety teams see risk before it turns into harm
For a long time, many Canadian safety leaders weren’t sure what role AI should play in protecting workers. Early conversations about AI in safety were often dominated by concerns about privacy, reliability, and whether algorithms could really keep up with chaotic factory floors or busy distribution centres.
Those questions haven’t disappeared, but the tone has changed. Today, organizations in manufacturing, logistics, and warehousing are far more likely to ask how AI-driven systems will integrate with their existing tools, what the data can actually tell them, and how they’ll demonstrate a return on investment to the C-suite.
“Canadian organizations have moved from exploration to expectation,” says Lindsay Martyn, enterprise sales director for Canada at Voxel. “They want technology that is deployable, auditable, and connected to outcomes, not just a buzzword checked off a digital transformation roadmap.”
Voxel, founded in 2020, is built to answer those expectations. The company’s industrial intelligence platform connects to existing camera infrastructure and uses AI-powered computer vision to scan live video feeds for emerging safety and operational risks. Instead of waiting for an incident to happen or reviewing a quarterly audit, safety teams see risky behaviours, near-misses, and hazardous conditions as they happen, with alerts and dashboards that show where to intervene first.
The idea grew out of a simple observation: most industrial sites already have cameras everywhere, but those cameras are only useful after something went awry. Voxel’s founders saw an opportunity to turn that infrastructure into an early-warning system that detects risk before it becomes an injury. Since then, the platform has scaled to hundreds of sites around the world, including some of the largest and most complex operations in high-risk industries.
Preventing danger before it happens
One incident from a customer facility illustrates the stakes. A powered industrial truck passed close enough to a pedestrian to knock her to the ground before the driver realized there had been contact. It was a serious near-miss that could easily have become catastrophic. With Voxel in place, the event was not only captured and alerted in real time; safety leaders also had detailed data on where, when, and how the near-miss occurred, giving them a clear path to redesign traffic flows and pedestrian routes before anyone was seriously hurt.
“What stays with me about that incident is that the pattern almost certainly existed before it became an injury,” Martyn says. “The technology doesn’t just record what happened; it tells you what’s about to happen early enough to do something about it. Our entire mission is to go from reactive to proactive with our insights.”
Those insights don’t stop at safety. By highlighting congestion points, workflow bottlenecks, and underutilized equipment, the platform often surfaces operational inefficiencies that have nothing to do with traditional injury metrics. Operations and industrial engineering teams use the same data to fine-tune layouts, scheduling, and material flow, turning connected safety tools into a driver of productivity as well as protection.
Designing safety around human action
Martyn believes that AI alone isn’t a silver bullet. One of the biggest challenges she sees is the gap between what the technology can surface and what organizations are ready to act on. If there isn’t a clear process for reviewing alerts, escalating issues, and feeding insights into coaching or engineering changes, even the best detection system can end up overwhelming teams.
“The lesson is to treat implementation as a behavioural design problem, not just a technical one,” she says. The most successful deployments are the ones where the human workflow has been designed as carefully as the AI application.
Looking ahead, Martyn expects AI safety tools to move from describing what’s happening on the floor to anticipating what’s likely to happen next, and from isolated site-level pilots to enterprise-wide intelligence that can be compared across dozens or hundreds of locations. For Canadian organizations under pressure to protect workers, demonstrate due diligence and keep operations running smoothly, that evolution could make connected, AI-driven safety less of an experiment and more of a foundational layer in how work gets done.
