Ireland’s multi-generations of dairy farmers know a thing or two about raising dairy cows. Its more than 18,000 dairy farmers tend 1.4 million animals and are recognized globally for productivity and quality. So, it’s no surprise that an Irish agtech company called Cainthus would invent a way to use artificial intelligence—the same technology developed for terrorist detection of humans—to manage dairy cows.
At its simplest, Cainthus’ technology has been described as facial recognition for cows, but Cainthus CEO Aidan Connolly explains that it is actually much more.
To be precise, Cainthus has developed a smart camera system that collects video data inside the dairy barn and uses artificial intelligence to uniquely identify and track behavior of all the cows in the barn. That information is used to develop key animal and farm performance indicators, which are delivered in the form of daily notifications and real-time detailed analytics to a dairy farmer’s phone. Such analytics help identify and analyze inefficiencies and animal health issues that need to be addressed to improve productivity and animal welfare.
The core dairy husbandry issues are the same, even “if you go back for the last 8,000 years of dairy farming,” notes Connolly. “Digital agriculture, for the first time, allows us to really precisely manage our cows, 24 hours a day, give them better welfare and make them more productive.”
The goal of facial recognition technology for humans has been much broader than simply identification and recognition. “It was designed to look at the overall way a person stands, the shadows they make, some of the other physical characteristics of that person. We’re using that same technology for cows,” said Connolly.
While milk production per cow is a metric that is fairly well tracked and measured on the farm, there are bigger questions of how to maintain those production levels. That is where farm management tools get less specific and are typically monitored by herd averages, rather than real-time data.
When it comes to an individual cow, Connolly explains, “We don’t know how much she eats, we don’t know how much she drinks, we don’t usually recognize lameness until it has pretty much already happened.”
The Cainthus vision system, and the resulting digitized data, changes that by gathering more information on a per-head basis to improve dairy farmers’ decision-making. For example, the Cainthus system might identify aggressive behavior by certain cows at the feeding trough or erratic feeding patterns among others.
Moreover, the information is in real time. Connolly shared the experience of one Canadian farm customer: The cameras showed that the animals were not drinking for more than a few seconds at a time. That led to the discovery of a stray electric current, undetectable to a human touching it, running through the water trough that caused the cows to back off. The problem was fixed, and, as a result, milk production rebounded more than two pounds the next day.
Likewise, early detection is particularly important for animal health management. “How much time does a veterinarian spend walking around the herd, trying to identify the cows that are in trouble?” asks Connolly. “We never get to stage one or stage two of laminitis (hoof inflammation that causes lameness). We recognize it at stage three or later. We don’t usually recognize mastitis (an inflammation of the udder) until we already have cell counts that end up affecting milk production.”
Cainthus’ technology adds value down the supply chain as well, addressing the growing interest in animal welfare, traceability and sustainability. Milk processors and bottlers, driven by consumer demand, are “starting to ask a series of questions of milk producers,” according to Connolly. “They want the number of lactations to increase, realizing the carbon footprint of a cow with more lactations is probably lower. They want to know whether the cows have had enough feed, enough water and other welfare concerns.”
Cainthus has plans to roll out its technology for other species, including beef cattle, hogs and poultry. Connolly noted that aquaculture also provides a “very exciting opportunity for the company’s technology. We don’t get to see (fish) because they’re underwater. Imagine what could be done if we start to detect, in real time, how effective our nutrition is, how effective our management is.”
For now, however, the company is focusing its virtual eye solely on dairy, which would seem to present some challenges, as milk prices and dairy income have seen a five-year economic decline. Connolly puts that in perspective: Dairy is a sector that is continually “boom-bust, boom-bust,” he says, which only serves to underscore the need for improving efficiency and productivity. He reports that progressive farmers are continuing to invest in technologies that make a difference.
Savvy investors are also investing in dairy’s future. Cargill, which has prioritized investments in what it calls digital insights, made a minority equity investment in Cainthus last year and maintains a strategic partnership for the rollout of the company’s technology, but everyone in the industry is paying attention.
In all, Connolly is very bullish on his cow-focused vision technology. “There have been very few unicorns in agtech so far. Only three or four have reached the billion-dollar valuation. Cainthus is going to be the next, in my view.”