Is voice input ready for farming?
It’s 2026 and accessible, high quality voice computing is finally starting to arrive. For an industry like agriculture, I think that matters a lot more than it might appear at first glance because it changes the way we interact with machines and, by extension, the way we interact with our own businesses.
In my view one of the main reasons digitisation in agriculture continues to lag behind where people would like it to be is user experience. More specifically, interface design that still feels stuck somewhere around 1995.
This is not unique to agriculture. Anyone who has spent time with enterprise or industrial software will recognise the patterns, ceefax was retired for a reason.in agriculture the problem is particularly visible because the environments in which the software must operate are so demanding.
Consider using a phone in a paddock. Bright sunlight washing out the screen. Gloves on your hands. Dust. Movement. Intermittent or completely absent connectivity. Under those conditions even well designed software becomes difficult to use. Poorly designed software becomes unusable and let’s face it most ag software is poorly designed.
Over the years a great deal of infrastructure has emerged designed to take the lumps and bumps out. Offline databases. Store and forward synchronisation. Caching strategies. Disconnected mobile application frameworks. Most of the building blocks now exist and many of them are mature, open source and widely used - just not necessarily so in ag. Crappy software engineering is common in ag.
The browser has also historically been somewhat constrained, not least because of Apple’s hobbling of Safari. But even there the tooling is steadily improving.
Then there is the cab of the tractor.
In-cab displays are crazy. The screens are filled with dense information, they have weak information architecture and pay very little attention to interaction design. They are often technically impressive but experientially exhausting. Interoperability challenges mean that many machines bring their own display, their own interface and their own mental model. A single cab can end up hosting several different screens, each competing for the operator’s attention.
All of which raises an obvious question. If screens are awkward in these environments, why are we so determined to keep using them as the primary interface? I think voice potentially offers another path.
The technology is no longer speculative. Models like OpenAI’s Whisper have made high quality speech recognition widely accessible and CHEAP. Apple now ships capable on-device models inside iOS. Nvidia and Google are pushing the state of the art further each year, while hosted services such as ElevenLabs provide excellent voice generation when connectivity is available.
Apple is even exploring more unusual approaches. The company recently acquired a startup called Q.ai, which has developed models capable of detecting tiny movements in facial muscles while someone speaks by analysing subtle shifts in the cheeks. It sounds sci fi, but technologies like this hint at a future where voice interaction becomes even more seamless, i think we will see this coming to AirPods.
Despite all of this progress, the agricultural technology sector is yet to produce many compelling examples of voice interfaces in practice. As usual, there is plenty of discussion, but relatively little execution.
That gap is partly what led me to start building a small mobile application at More Than Machines called Bilby.
Bilby is deliberately simple. It is essentially a voice memo system designed for producers to capture what they are doing as they move through the day. Spray applications. Livestock movements. Observations in the paddock. The small pieces of information that rarely make it into a formal system because recording them is too awkward at the moment they occur.
Instead of navigating menus or filling in forms, the producer simply speaks.
“Just moved the steers into the back paddock. About fifty head. Water trough looks a bit low.”
Or perhaps while driving a spray rig.
“Started spraying canola in the north block. Using clethodim. Wind is picking up a little.”
Or a quick observation while walking through a crop.
“Seeing a bit of rust starting along the eastern fence line.”
Bilby is intended to be the entry point for data flowing into the More Than Machines farm memory stack. Voice notes are captured, structured and turned into events that can be stored, analysed and eventually used by other tools such as farm management systems and of course agents.
The idea is straightforward. Reduce the friction of capturing information and you change the amount and quality of data that is available later.
Voice won’t replace all interfaces in agriculture. Screens will always be important and there are many tasks where visual interaction is essential. But for the simple act of recording what is happening in the real world, voice and sound may turn out to be the most natural interface we have and agriculture, with its harsh environments and busy operators, may turn out to be one of the places where voice computing really provides significant value.