When farmer Frank installed Halter on his phone, he did something that would have previously been unimaginable for farmers: he moved his entire family to France, whilst continuing to manage his farm remotely with complete confidence.
This freedom to run a livestock business remotely is possible thanks to New Zealand-founded agritech company Halter, which has revolutionised livestock management with connected collar technology, built on Amazon Web Services (AWS) from day one.
By enabling virtual geographic boundary setting for cattle fences via mobile app, while retrieving data on animal wellbeing and behaviour, farmers can make more informed grazing decisions that turn into better farming outcomes.
“With Halter, farmers no longer need to be out on the farm at 4am to move animals. They can manage their operations remotely and spend more time focusing on their business and their families,” said Kurt McAlpine, cloud architect at Halter.
Now, AI is fast-tracking the next phase of farming evolution. Leveraging Amazon Bedrock and Anthropic Claude, Halter has deployed an AI agent to help its engineering team delegate workloads and improve operational efficiency across farms, providing a more seamless and reliable digital experience for farmers than ever.
A reliable system for scalable, connected farming
For Halter, AI is not an end goal, but an enabler that helps farmers spend less time on administrative and operational tasks, and more time focusing on the strategic decisions that improve farming outcomes, such as livestock health and pasture conditions.
Behind every farmer's ability to herd cattle with the press of a button lies a sophisticated engineering challenge: keeping Halter's platform running reliably at scale, across multiple countries and time zones.
To address this, Halter built Clank, an AI agent built on Amazon Bedrock using Anthropic Claude that automates incident response and routine engineering workflows, without sacrificing the security and reliability that safeguards farmers’ livelihoods.
"There's an animal on the other end of that technology, so it's very important to us that everything is robust and scalable,” said McAlpine. “We chose Amazon Bedrock as it gives us the reliability and data privacy we need, to be secure and confident in deploying Clank.”
Building Clank on Amazon Bedrock
Halter has over 16,000 Amazon CloudWatch alerts in place, an intelligent observability service from AWS that monitors every component of Halter’s system and identifies application issues before they impact service to the mobile app.
Historically, Halter's engineering team would respond to more than 40 alert tickets each day when farmers needed technical app support, taking manual time to investigate alerts, identify root causes, and determine appropriate actions.
With Clank, the agent can automatically analyse alerts, perform root-cause investigations, and prepare fixes upfront for Halter’s engineers to review and action. Clank is also integrated into Halter’s existing workflows through tools such as Slack, GitHub, and Linear, meaning it constantly listens for production alerts and user requests.
For example, when an Amazon Cloudwatch alert is triggered, the system initiates an automated workflow, retrieves relevant logs and system data, then connects to Claude by Anthropic in Amazon Bedrock to analyse the issue.
Even in the middle of the night, Clank then generates a code change before an engineer begins their workday, allowing teams to review and approve changes rather than start from scratch.
“On an average day, Clank works the equivalent of 231 hours,” said Antony Southworth, lead data engineer at Halter. “That’s nearly 29 times a typical engineer's workday - at an inference cost of approximately US$30 per hour, making AI-driven incident response both faster and significantly more cost-effective than manual triage.”
Saving 215 hours on manual workflows
In the first few months of deployment, Halter saved more than 215 hours of engineering time through AI-driven investigations and the automation of over 90 weekly tasks via Clank.
Clank has also accelerated faster resolution of platform issues, with an AI-generated code change now reviewed and embedded into the system in just 48 minutes, with over 2,500 code changes processed over an 11-week period.
By reducing the time engineers spend responding to incidents, Clank helps ensure the platform
remains reliable and frees up engineering time for more valuable efforts - developing new model capabilities that enhance the digital experience for farmers, allowing farmers to manage herds productively and with confidence.
“AI is completely reshaping what's possible for our industry, and we at Halter are proud to support farmers through this next phase of evolution to make their lives easier than ever,” said McAlpine. “AWS have helped us to transform into an AI-led business, skyrocketing our opportunities to scale our farming technology to the world and be a leading key agricultural exporter for New Zealand.”
"Across New Zealand, we're seeing organisations embrace AI to solve real-world challenges, improve productivity, and create new opportunities for growth," said Manuel Bohnet, Country Manager, New Zealand at AWS. "Companies like Halter demonstrate how cloud and generative AI can help agribusinesses innovate faster, operate more efficiently, and build leading, globally competitive businesses - right from the heart of New Zealand.”
