[yunnpp]

I'm not the expert and the details that the company would put out are obviously obtuse, but: image detection/identification, predictive policing, and planning. The latter sounds to me like they'd have some system where people enter reports in natural language and an LLM assembles the information together and then proposes some plan of action. As opposed to having to have a more structured data entry and the friction that comes with it. It's all in the article if you read between the lines, really.