Empowering Vulcan Steel - Health & Safety Case Study.
Microsoft NZ partner Datacom worked with Vulcan Steel to build a workplace safety solution using Microsoft AI and machine learning.
Vulcan Steel makes about 3,000 deliveries of steel a day to businesses throughout New Zealand and Australia – which means that each day, its employees need to use their training to figure out how to safely get large, heavy and unwieldly pieces of steel off of its trucks and into the hands of a very diverse group of customers.
“It’s an awkward product to transport, and it’s difficult to design out all of the risks,” said James Wells, who acts as the company’s chief information officer. “So essentially what that means for us is one of the key requirements or skills for us to keep people safe is around education.”
For years, Vulcan Steel did what most companies do – they educated their employees about safety before sending them into the field, and then they did additional training as needed if someone reported an accident or near miss.
Now, they’re using artificial intelligence to try to more proactively prevent accidents and near misses before they happen. The company recently started using Microsoft Cognitive Service’s Custom Vision tools to evaluate camera footage from the company’s trucks for actions that could be risky or lead to an accident.
The computer vision tools are able to do what the human eye couldn’t reasonably do – sift through thousands of pieces of footage a day to look for potential risks – freeing up the company’s workers to review just a small subset of footage that has been flagged as a possible concern.
That, in turn, is allowing Vulcan Steel to focus its education efforts on what it sees as the most worrisome or risky scenarios. Wells said accidents were already exceedingly rare, so the goal is to build the company’s culture of safety.
“What we’re hoping is we will measure the number of education discussions that take place as a result,” Wells said. “From our point of view, if we add an additional number of safety discussions to our organization, there’s not really any negative that can come of that.”
Vulcan Steel doesn’t have a large staff of developers, and it certainly doesn’t have a team of AI experts. Wells said the development of this AI-based system was basically the work on one enthusiastic .NET developer who saw the potential for how AI could help the business.
