More important, however, is the urgency of such a tool, he said. Sustainability teams are being stretched more than ever. With narrowing resources available, they’re being asked to make their work relevant to sourcing, procurement, supply chain operations, even product design.

“They’re getting pushed to come together with all of those teams, and that’s exactly what this product does,” Vranes said. “It serves all of those teams and, at the same time, it gives a single risk picture and an opportunity picture to all of them. Because the sustainability team has its own lens and its own risks and opportunities to look at. And it’s the same thing on the sourcing side, the same thing on the procurement side, the supply chain operation side and the product design side. All of them have their own kind of unique needs in this space, and this product helps all of them at the same time.”

Take climate change and the warming trends in already heat-stressed areas, for instance.

“So say I’m in an area that regularly sees 35-or 40-degree Celsius days, and now I’m turning that up into 42- and 45-degree Celsius days, those are heat stress risks that become labor and human rights risks,” Vranes said.  “The climate models tell me this region is getting hotter in this way, the weather models are telling me that maybe next quarter I’m going to start seeing extra heat waves here, and then my own internal data about my supply chain is telling me this is where I have operations—Tier one here and Tier two here—and this is where I might place vulnerability that leads to both workers rights issues, but also absenteeism. I might have fewer workers showing up for work because it’s too hot.”

Take that concept further and companies can take on what he calls “multivariate scenario planning” by adding predictive insights to provide more context and clarity around hotspots that can enable smarter, snappier sourcing decisions by optimizing for anything from decarbonization for the greatest impact to shifting energy sources to minimizing logistics costs. In the longer term—like its Higg counterparts, the tool will be an iterative one—Worldly Axion could anticipate problems before they arise.

“You might have a strike at a port in two months because the workers are agitating, and so we’re anticipating, based on news reports and social media, that maybe the Hong Kong port is going to shut down,” Vranes said. “Well, you’d better start thinking ahead and start moving your production around.”

While today’s AI bots are underpinned by complex mathematical models that allow them to make sense of vast amounts of data, critics also say they’re prone to making stuff up, a phenomenon known as hallucination. Nearly three years after the advent of ChatGPT and the embedding of AI into daily life, erroneous information and sometimes outright fabrications are still a common occurrence. With so much at stake, what’s to say that Worldly Axion won’t fall into the same morass?

Vranes said that models are capable of knowing when they’re hallucinating or when they’re likely to hallucinate. Those that prioritize speed over accuracy might spit out plenty of seemingly unrealistic untruths. Other, more thinking-focused ones are taught to dig deeper, even question themselves, to provide answers with higher degrees of confidence. The chatbot that’s built into Worldly Axion, similar to the Worldly assistant that’s already on the core platform, he said, is the latter.

“That model is really good at just saying, ‘I don’t know,’” Vranes said. “‘You asked me a question, but I cannot answer it’ or ‘I don’t know how to answer this question because I don’t have the information.’ And so we’ve seen really good results with that, too.”

AI has also stirred up controversy for its hefty environmental footprint. According to one recent estimate, ChatGPT guzzles nearly 10 times as much power as a normal Google search. At the same time, one study by the Boston Consulting Group—commissioned by Google—found that artificial machine learning, if applied prudently, could reduce greenhouse gas emissions by 5 to 10 percent by 2030. Worldly, for its part, expects its technology to get increasingly more efficient. Vranes also said that text-based AI is less energy-intensive than systems that spit out videos and images.

“We have a kind of stance in our technology team of looking at how to optimize for energy and so how to not always just go, every time there needs to be an AI query, ‘Don’t even think about the energy use. Just go,’” he said. “There are always techniques to say, like, well, do I really need to use this kind of model instead of this model?”

For now, Worldly Axion isn’t even at step one. It’s more like step zero, Vranes said. But things are going to evolve very quickly, he added.

“I think the bottom-line summary of what Worldly Axion does is it transforms sustainability data and risk data, both of which are overwhelming data sets that are not connected in almost any organization, and we’re transforming them into what we would call clear, prioritized actions,” Vranes said. “And for what reason? For more resilient supply chains, for better investment decisions, for better operational decisions.”