Halter launches Mav to answer dairy farmers’ questions using collar data
Mav combines information on grazing, pasture, animal behaviour, health and reproduction, allowing farmers to query the thousands of farm-performance data points available to them each day. Users can request a morning herd assessment, identify animals requiring attention or compare current pasture growth with earlier seasons. They can also arrange for a recurring question to generate an update each day. During beta testing, users submitted more than 100,000 questions.
Craig Piggott, Halter’s founder and chief executive, described the assistant as a way to make the information collected by its technology more useful for farm decisions. He said the company interprets AI as Agricultural Intelligence and has spent years developing tools that provide multiple layers of farm information. In his view, helping farmers make more of their hundreds of daily decisions well, and sooner, could increase grass production, pregnancies and milk output.
As an example, Halter described a large dairy operation where about 50 cows might enter heat in one day during the mating season. Its existing system identifies those animals, after which farmers assess each cow’s previous heats and the timing and intensity of the current heat. According to the company, Mav can examine those factors together to help farmers determine which animals need closer inspection and when.
The assistant uses an anonymised, aggregated collection of Halter data that the company considers among the world’s largest datasets combining pasture and cattle information. Its coverage includes a cumulative 10,000 breeding seasons, seven million detected heats, 1.2 million health cases and over 500 million days of collar use on cattle. The company’s collars have produced more than 1.2 trillion records.
Halter has also opened its Agricultural Intelligence Archive, describing it as the first physical archive devoted to pasture-farming insights. Located at the company’s research and development farm in Morrinsville, Waikato, the temperature-controlled facility occupies a cattle underpass built into a grass-covered hillside. It will preserve foundational farming knowledge across seasons, with findings from Mav verified before being added as physical records. Seed banks inspired the project, which Halter intends to use to retain and share knowledge derived from farms, animals and successive seasons. Piggott said the ability to learn from farms at scale gave the company a reason to preserve some of those findings.
The archive’s inaugural book, Agricultural Intelligence Volume 1, covers calving and the period before mating in 2026. It draws on aggregated information from New Zealand dairy farms involving about 800,000 cows wearing Halter collars; the company has two million collared cows worldwide. Topics include the acceleration of pasture growth between winter and spring in 2026 and what behavioural records from nearly 590,000 cows indicate about recovery following calving. Five New Zealand farmers contributed their farms’ data for case studies describing their use of Mav.
Wider findings preserved in the archive will combine data and remove identifying information. Printed copies of the first volume will be available for viewing, and the volume can be accessed online free of charge. The collection also includes historical logbooks donated by New Zealand dairy farms and documentation of other physical farming technologies. Halter expects the collection to expand and plans to admit visitors on designated open days.




