Avalanche risk and its management are unknown to me, but not to ski patrols, experienced backcountry trekkers, scientists, and national weather services of many countries, etc. I saw no evidence of trials to test to see how the technology performs against real conditions.
If you can’t trust this product’s recommendations within some margin of safety then it’s value proposition disappears.
Why not market it instead as an experiment and seek some way of collecting data to improve the process or develop a model fine tuned to the inputs?