Data & Analytics

When we talk to safety leaders about computer vision AI, the conversations are almost always positive. They understand the technology, they see the value, and they're keen to explore it. But when it comes to rolling the technology out to the wider workforce, a different set of conversations begins. Workers have questions. Concerns. And sometimes, deeply held beliefs about what AI cameras do that simply aren't accurate. These misconceptions aren't unreasonable. Headlines about

If you run safety across multiple sites, you've almost certainly sat in a meeting where someone compared Lost Time Injury Frequency Rate (LTIFR) across locations and drew conclusions about which site is "doing well" and which one "needs attention." On the surface, it seems like a reasonable exercise. In practice, it's one of the most misleading things you can do with safety data. The problem isn't that benchmarking is a bad idea. Benchmarking is essential for multi-site organ

Here's a pattern I see all the time. An organisation invests in safety technology. The system starts detecting events. The dashboard fills up with data. And then... nothing much changes. Not because the technology doesn't work. It works fine. The detection is accurate, the data is rich, and the dashboard looks impressive. The problem is that nobody has built the workflow between "we detected something" and "we did something about it." The data sits in the dashboard, gets revi

If I asked you to point to the most dangerous spot in your facility right now, you'd probably have a pretty good answer. The intersection near the loading dock. The blind corner behind Rack 12. The pedestrian crossing by the dispatch area. But would you be right? And more importantly, would you know whether it's more dangerous at 6am than at 2pm, whether the risk shifted after you rearranged the racking last month, or whether your third-party delivery drivers create a differe

If I asked you which shift at your site has the highest safety risk, could you answer with confidence? Most safety managers have a gut feel. "Nights are worse." "The Friday afternoon shift is a bit loose." "Changeover is always hectic." But gut feel is not data. And without data, you're managing shift risk with intuition rather than evidence. When organisations deploy computer vision AI and start capturing safety events continuously across every shift, one of the first thing

Here's something that should bother every safety manager: you almost certainly know about the incidents that went wrong. The ones that resulted in an injury, a lost time event, a workers' compensation claim. Those get recorded because they have to be. But what about the hundreds of events that nearly went wrong? The forklift that turned a blind corner at speed, missing a pedestrian by a metre. The pallet that slipped off a rack but landed in an empty aisle. The worker who ste

You know your safety data matters. You track it, analyse it, and use it every day to make decisions that keep people safe. But the moment you walk into a board meeting and start talking about TRIR, DART rates, and near-miss frequency, you can see the eyes glaze over. It's not that your board doesn't care about safety. Most directors take their governance obligations seriously, particularly in Australia and New Zealand where officers have personal liability under WHS and HSWA

Ask most EHS managers what they measure, and you'll get the same list: lost-time injury rate, total recordable incident rate, workers' compensation claims, days away from work. These are all lagging indicators, and they're all telling you the same thing: what already went wrong. There's nothing wrong with tracking them. You need lagging indicators for compliance reporting, benchmarking, and understanding the severity of incidents after they occur. But if lagging indicators ar