An AI system that makes city waste collection smarter: it monitors bin fill levels, predicts when they will overflow, and dynamically reshuffles collection routes. Built by a Kenyan team with Fabric Data Agents, Copilot Studio and Azure OpenAI.
The name of the award describes the axis it was judged on: not novelty of technique but whether the application holds together. Waste collection is unusually well behaved that way, since fill levels are a real sensor signal, a forecast becomes a changed route, and the benefit shows up as kilometres not driven. The stack fits the same way, each sponsor product doing a distinct job instead of overlapping for show.
Summaries are written by this site. Project pages include demo videos, screenshots and the team's own write-up (videos may autoplay).