Posts

What forecasting delivery risk looks like in practice, and how far ahead a model can reasonably see on a live route.

A short case study on re-sequencing a live route around a late pickup while keeping every promised delivery window.

Traffic is not the expensive one. A field ranking of the disruptions that actually turn into missed windows, and what each one costs.

Separating the demos from the daily reality: where models genuinely help a dispatcher, and where a human still has to own the call.

Changing a route at 11am is easy. Doing it without knocking three other customer windows out of the plan is the hard part.

A closer look at the judgment calls a dispatcher makes in a day, and why most of them are made with incomplete information.

Every dispatcher makes dozens of judgment calls a day, most of them with incomplete information and a clock running. What the research says about deciding well under pressure.