Datamole

The deepest published work is all one client: Lely's milking-routine personalization calculates an optimal treatment per cow on every robot visit, fed by Datamole's IIoT platform collecting from more than 30,000 robots worldwide across 5 million daily milking sessions, and the write-up reports up to 80 percent fewer non-optimal milkings and 0.2 kg more milk per robot-minute.

The Lely Service Support App is the second full write-up and scaled from one service center to more than 200 centers and 2,000 regular users over eight years of collaboration, though it is service analytics and an app rather than a model in the loop.

Beyond those, the case index is two-sentence summaries—Agrifac sprayer telemetry, Cup&Cino device data, Thermo Fisher Scientific microscope condition monitoring—and the third full page, Calf2Cow, is a Eurostars-2 and Horizon 2020 co-funded project run with Triodor and Lely rather than a client engagement, so ask for the list behind the "100+ successful projects" and ">80% of projects deployed" figures.

The weaker fit is a buyer shopping for generative or agentic AI—nothing published describes LLM work, and the visible depth is industrial IoT, predictive models and the applications built on top of them.