Farms generate a huge amount of information. Some of it is written down. Some lives in spreadsheets. Some sits in emails or text messages. Some is photos on a phone. Some is simply remembered. And much of it, over time, gets lost.
This isn't because farmers don't value data - it's because recording it often gets in the way of getting the job done. When you're mid-drench, shifting stock in the rain, or troubleshooting a pump at the far end of the property, pulling out a laptop to update a spreadsheet isn't realistic.
Technology in this space has often tried to solve the problem by adding more structure - more required fields, more categories, more complexity. But this creates friction. If recording information is difficult, it simply doesn't happen.
At AgRhythm, we believe farm record-keeping should be simple, flexible, and useful from day one. That's why we've built AgRiver.
A Digital Farm Journal, Not Just Another App
AgRiver is designed as a digital farm journal. The idea is deliberately simple: make it easy to capture what's happening on the farm, in whatever form makes sense at the time.
That might be a quick note. A photo. A short video. A voice memo. A paddock observation. A stock movement. A weather-related note. A task or event. The goal isn't to force structure too early - it's to capture information while it's fresh and meaningful.
The best time to record something is right when it happens. The best format is whatever the farmer has to hand. AgRiver is built around that reality.
We've also built Barry - a conversational farm guide available at barry.agrhythm.nz - as a natural way to interact with this record. Instead of filling out forms, you can simply talk to Barry about what's happening on the farm. He asks helpful questions, remembers the details, and builds the record through conversation. Record-keeping that feels like a yarn, not admin.
Building a Farm Data Moat
There's a concept in technology called a "data moat" - a body of information that becomes increasingly valuable over time. On a farm, this might look like seasonal pasture observations, infrastructure changes, stock performance notes, weather impacts, management decisions, and their outcomes tracked over months and years.
Individually, each entry may be small. But together, they create context. And context is where real insight begins to emerge.
This is where AgRiver becomes more than a digital notebook. Every note, photo, and observation adds to a growing picture of the farm - one that becomes richer and more useful with each season. In New Zealand pastoral farming, where so much depends on understanding how pasture, stock, and weather interact across specific paddocks over time, this kind of accumulated knowledge is genuinely powerful.
The most valuable farm data isn't what's recorded once - it's what's recorded consistently, over time, and connected together.
AI-Powered Farm Insights
As the record grows, AgRiver uses large language models to help make sense of the data. This allows the platform to summarise recent activity, highlight patterns over time, connect related events, surface useful observations, and assist with planning and decision-making.
Instead of just storing information, the platform starts working alongside the farmer. For example:
- "Over the past three seasons, this paddock has dried out earlier than the rest of the block - worth considering for earlier grazing rotation."
- "Water trough issues were recorded three times in the back block last summer - might be worth a permanent fix before next season."
- "Stock performance improved after the management changes you recorded in early spring."
These are the kinds of insights that are nearly impossible to track manually across a busy season, but powerful when surfaced at the right time. The goal isn't to replace decision-making - it's to make the farmer's own observations work harder for them.
Designed to Be Flexible
AgRiver isn't limited to one type of data. It can incorporate farmer-entered notes, photos and videos, drone imagery, sensor integrations, and connections to existing farm management systems. This flexibility matters because every farm operates differently and every farmer records information differently.
Under the hood, AgRiver is built on our geo-temporal data platform - meaning every piece of information is tied to a location and a point in time. A note about a paddock isn't just a note - it's connected to that paddock's history, its drone imagery, its pasture data, and everything else recorded there. This spatial and temporal context is what allows the AI to surface insights that a flat list of notes never could.
AgRiver adapts to how the farmer already works, rather than forcing a rigid workflow. The technology should fit the farm - not the other way around.
Building With Farmers, Not Just For Farmers
Right now, AgRiver is at an early stage. We've built a working platform and we're actively using it ourselves. But the real value comes from working alongside farmers and learning what actually works in day-to-day farm life.
We're currently looking for farmers who are interested in collaborating with us - not as customers, but as partners in shaping the platform. We're particularly interested in learning what information farmers naturally record, what they wish they had recorded, what feels useful, what feels like friction, and how technology can fit naturally into farm routines.
A Long-Term Vision
Over time, we believe this kind of platform becomes increasingly valuable. As more information is captured, insights improve, context deepens, patterns become clearer, and decision-making strengthens. This creates a record that serves the farm over the long term.
Not data for the sake of data - but information that helps farmers understand their land, their decisions, and their outcomes more clearly. Combined with drone imagery calibrated to New Zealand conditions and conversational AI through Barry, AgRiver is designed to be the kind of tool that gets more useful the longer you use it.
Because the best farm insights don't come from technology alone. They come from farmers. We're just building tools to help capture and support that intelligence.