Data ingestion pipeline
A pipeline that accepts time-series behaviour data, normalises it per animal and stores it for analysis. Built and running against sample datasets.
Technology
Per-animal baselines, not herd averages.
Our intended approach in three layers. Each item below states honestly whether it exists today, is under development, or is still on the roadmap.
The intended pipeline
01
Collect
Behaviour and location data per animal, at intervals frequent enough to see change rather than averages.
02
Model
Establish each animal's own baseline, then measure deviation against it instead of a herd-wide norm.
03
Surface
Present the small number of animals worth a second look, with the evidence behind the flag.
Capability status
System architecture
The overall design — ingestion, modelling, records and interface boundaries — is defined and documented internally.
Behaviour signal models
Models that flag deviations in movement, feeding and rest patterns. In active development and internal testing; not validated on live herds.
Grazing & movement analysis
Prototype analysis of where a herd spends its time and how that shifts over days and seasons. Not field-tested.
Per-animal records
A timeline per animal — treatments, weights, births and movements — planned but not yet built.
Alerting & mobile access
Real-time alerts and a field-friendly mobile interface are future goals, dependent on validated models.
No HerdsenseAI hardware is deployed on any farm. Nothing on this page describes a certified, commercially available or independently validated product.