CattleEye AI Reduces Lameness 10% Boosts Cow Lifespan 40%

Terry Canning, who grew up on a small dairy farm in County Armagh, Northern Ireland, initially pursued an academic path in engineering after graduating. However, in 2004, at the age of 30, he returned to his family home to establish his own company, focusing on developing a software product for managing livestock. This decision was driven by the importance of animal analytics in the agricultural industry, particularly in light of recent outbreaks of foot-and-mouth disease, which highlighted the risks of not keeping records of animal movements.

Canning's initial idea was to attach devices to animals, akin to a "Fitbit for cows," but this approach proved to be expensive and impractical. Batteries would run out, and ear tags could get caught and removed as cows moved through the farm. With the help of his cofounder, Adam Askew, a researcher in computer vision for medical purposes, Canning found a more effective solution. They devised a system using computer vision to analyze the health of cows, particularly whether they were limping. This technology was crucial in addressing lameness, a significant issue in many dairies, where approximately 30% of cows do not walk as well as they should. This problem has both economic and environmental impacts, as immobile cows are culled, leading to increased resource use and greenhouse gas emissions.

The company, CattleEye, was founded in 2019 with a core technology that has remained largely unchanged. An AI-powered camera, positioned above the route dairy cows take daily, identifies each cow based on its unique markings and assesses its gait. The camera identifies around 16 points on each animal, such as the ear, the back of the head, and the spine, using the shifting of these data points to analyze the cow’s gait. CattleEye’s AI is composed of advanced deep learning models trained on data generated by veterinarians, with a trove of around 100,000 videos of cows annotated and scored by veterinarians for observed health.

When in use, the AI gives each cow a score between 0 and 100. Any cow scored over 50 is deemed lame and worthy of a checkup by the farmer, who accesses the data through a web app. CattleEye can identify lame cows four weeks earlier than most humans, thanks to its machine-learning-powered analysis. The system typically costs farmers about $1.45 per cow per month but can deliver "at least a 10x return." It also slashes lameness by roughly 10 percentage points and cuts emission intensity by up to 40% by increasing the lifespan of cows. The science behind CattleEye has been evaluated by independent scientific studies, which came to similar conclusions.

The setup costs are kept low by using off-the-shelf products, avoiding high upfront charges that could scare off farmers. The AI system is a patented and complicated process, but the hardware involved consists of a simple, mass-produced camera posted at least four meters above the ground where the cows walk, as well as an Ethernet switch and an SD card. The system works even if multiple animals walk underneath the camera at once, accommodating the natural behavior of cows.

One of CattleEye’s earliest commercial customers was a U.K. supermarket chain. The company has been working with CattleEye since 2019, using the technology to ensure that their customers know their farmers are fairly treated and their cows are well cared for. This early adoption helped CattleEye secure seed funding of £500,000 from a venture capital fund. Further funding included a $2.5 million round in 2021. The company launched a pilot project in 2020 at three dairy farms, where a veterinarian would rate the cows’ health over time as the cameras watched, providing ground truth data for the AI system.

Today, CattleEye is a digital custodian for more than 200,000 cows worldwide, with an increasing customer base in the U.S. after being acquired by a food technology supplier in March 2024. Clients include global dairy giants and food groups. The company is now exploring other uses for its technology, having completed a pilot project with poultry to measure their well-being. Canning envisions a future where the combined expertise of GEA’s data and CattleEye’s AI could provide specific advice on how and when it’s best to milk a particular cow, as well as insights into when a cow is calving and how well they’re eating. This technology aims to bring about a better planet, feed more people, and provide more nutrition in a more sustainable way.