Moving from the clipboard to the Cloud

Harnessing AI-driven insights for precision livestock management in the dairy sector

By Nial O’Boyle

In medicine, artificial intelligence (AI) has revolutionised tumour detection by identifying minute anomalies in medical scans, often before symptoms emerge. 

Deep learning models now outperform clinicians in spotting early-stage tumours, not by replacing expert judgement but by providing high-frequency, consistent, and fatigue-free analysis. It’s not necessarily in the identification when challenged side-by-side, but the ability for AI to be consistent, with a frequency and durability that a human cannot match. 

Taking it to the dairy parlour

CattleEye’s founders CEO Terry Canning and CTO Adam Askew had a perfect blend of experience in cloud-based herd management software, and running this type of AI-enhanced tumour detection. They were able to attract several scientists that had experience with this technology, applying similar principles to a very different problem: lameness in dairy cows. 

Mobility scoring at scale, with the frequency and consistency needed is extremely difficult. Utilising the same principles that augmented tumour detection, CattleEye has enabled cows to have mobility scoring with the same expertise as a trained veterinarian, but daily instead of monthly or quarterly.

As with the tumour detection, you can identify problems early, enabling interventions that improve outcomes and reduce long-term damage. When every cow is scored daily, herd managers can move beyond population averages and focus on individual condition trajectories; detecting outliers, adjusting rations quicker, and responding before performance suffers. Additionally, unlike manual scoring, which becomes impractical on large units, automated Body Condition Scoring (BCS) scales effortlessly across herds of 1,000+ cows, without adding to labour demands

Augmenting BCS

BCS is taking advantage of the same principles to offer novel insights to manage dairy cows. BCS is a visual or tactile assessment of fat reserves, typically on a 1–5 scale. Experienced scorers palpate and observe key areas (rump, tailhead, loin) to assign scores. 

However, manual scoring is labour-intensive and subjective, leading to inconsistencies. Trained observers often disagree, and scoring errors are common. In practice, it is difficult to score large herds frequently, as a result many have limited BCS data, hindering precise management of nutrition and health. 

CattleEye solves this problem. Siachos et al. (2024) from the University of Liverpool validated the CattleEye BCS module. Against human scores (9,657 paired observations), the AI had substantial agreement (weighted κ ≈0.69) with experts, with 84.6% of scores within ±0.25 BCS and 94.8% within ±0.5 

Augmenting humans

Augmenting the skills of a human by leveraging AI tools like CattleEye provides frequent, unbiased estimates of individual cows. Instead of relying on a handful of condition scores for each lactation, farmers now have access to hundreds of data points per animal, transforming BCS from a periodic snapshot into a continuous signal. This enables consistent, herd-wide monitoring that supports optimised feeding, fertility management, and early detection of at-risk animals. 

Critically, it allows large herds to be managed with the same attentiveness as small ones, without the labour burden. As with tumour detection in radiology, the real shift lies not in replacing human judgement, but in delivering the right information, at the right time, to act.

To find out more about how AI can support animal welfare on your farm, please contact the team on contact@cattleeye.com