An AI cow-collar that’s helping raise yield for farmers, dairies

Last September, Mahesh Gangurde was sifting through insurance papers in his office in Nashik when his phone beeped. The alert said four cows on his farm 80 km away in Chandwad were showing reduced rumination. He had lost three cows to delayed detection of diseases in 2024, so Gangurde wasted no time and called up his veterinarian.

The cows were diagnosed with the onset of protozoan disease, the same one that had killed the previous lot. They were put on medication and within six days, they were back to normal without much loss of milk.

Dr Pravin Gavade was attending to patients in his clinic when he got an alert that four of his cows were in silent heat. The alert specified the best period of insemination for each. Three of them are six months pregnant now.

Timely alerts to Zuhaib Pasha in Chikkaballapur in Karnataka helped bring down his veterinarian costs for 15 cows from Rs 11,000 per month to Rs 2,500, and his milk production has stabilised.

All these dairy farmers have in common a collar equipped with sensors that monitor cattle health, provide alerts in case of abnormalities, and indicate the optimal time for insemination. The collar, developed by Pune-based Areete Business Solutions (ABS) and incubated at BAIF Development Research Foundation, has attracted the attention of major dairy industry players, including Amul and Chitale.

ABS is an agri tech firm founded by Srinivas Subramanian and V S Shridhar. Subramanian had worked with Aditya Birla Group, Bajaj Auto, and Crompton Greaves, and Shridhar with IBM and Tata Group.

Shridhar says they began looking at ideas for farmers in 2016, and zeroed in on cows as they are a source of constant income for farmers, unlike weather-dependent agriculture. “Automations are available for everything after milking the cow, but there is nothing for the animal. India, despite being the largest producer of milk, has a severe shortage of veterinary doctors. The onus is on the farmer to identify issues with the cow, or risk losing it to illness,” he says.

Mahesh Gangurde started using the collar in 2024 after three of his cows died. “The doctors told me that if we had brought the cows to them earlier, they could have saved them.” He went to bigger dairies to understand how they detect diseases, found some sensor-based collars, but found suppliers only sell in bulk—to farms with more than 100-150 cows.

And then he found ABS. He tested them on his best milk-producing cows, and when he found they worked, he bought for all his 25 cows.

The collar uses four types of sensors—gyroscopes, accelerometer, magnetometer and temperature sensor. They collect 550 data points in the cow, everything from its movement to rumination to health and behaviour. This information is sent to a database, where AI/ML is applied for analysis. If there’s an issue, it sends an alert on an app the farmer has on his phone. The system can communicate in nine Indian languages. Since some farmers may miss notifications, ABS employs care managers who call farmers, especially for heat alerts.

Pravin Gavade, who has been using it for the past eight months for his 40 cows (Holstein Friesian and Desi breeds), says the detection of silent heat in cows is the best feature of the app. “If the cow doesn’t conceive within a particular time period after calving, then we have to wait a long period to milk her again as lactation happens only after calving,” he says.

Chitale Dairy, one of the oldest organised private dairies in the country, collects 10 lakh litres of milk daily through a network of more than 100 milk collection centres from farmers in Sangli, Satara and Kolhapur. Vishwas Chitale, CEO & CTO of the dairy, says since they started using Areete’s collar, they’ve seen a 10% increase in conception rates. “The system has also consistently delivered nearly 93% accuracy in heat prediction, setting a new benchmark in precision and reliability,” Chitale says.

ABS collars are much more affordable than imported ones, and work on both cows and buffaloes. Once attached to a cow, Shridhar says, it takes two weeks to capture the cow’s movements and other parameters and standardise it to arrive at a baseline for each cow. “Our tech predicts heat in cows with 95% accuracy, and health issues with 90% accuracy,” he says.