Why Halter runs its support agents on Claude's top AI models | The AI Corner
A Kiwi farming tech company is getting pretty serious about AI, building smarter support agents that actually know their stuff instead of just throwing chatbots at the problem.
Key takeaway When your agritech platform talks to farmers, you'd better make sure it's talking sense.
Halter, the New Zealand agritech outfit, has brought Pouria Asghari, Head of GTM Ecosystems & AI, in to explain why they've chosen to run their support agents on Claude's top-tier AI models. The company is deploying Claude alongside RAG (Retrieval-Augmented Generation) pipelines - basically teaching AI to pull the right information before it answers - to handle both customer support and internal business automation.
The setup means Halter's support agents can dig into actual product knowledge and farm-specific context before responding to farmers, rather than guessing or serving up generic answers. That's a meaningful shift for an agritech company where getting the detail right actually matters on the ground.
For the Kiwi tech scene, it's a solid case study: a homegrown company choosing its AI infrastructure carefully, investing in proper RAG architecture, and thinking through how AI fits into real customer relationships - not just bolting it on because it's trendy.
“Turns out farmers need their AI to know the difference between a cow and a blade of grass - who knew?” — The Nation's Mother, The 6 O’Clock News