Can you dig it? The AgTech promise of robotics - but will the robot farmer please stand up?

Agricultural technology – aka AgTech or agritech – is a critical sector for any nation that has significant commercial farming and food production industries, both arable and pastoral. So, how can robotics and AI help? The answer is that the theoretical big-picture wins of AgTech are easy to state, but the real-world challenges are sometimes harder to back, meet, and solve, according to a Westminster policy conference.

Specialist robots could help harvest crops faster, more efficiently, more reliably, and at greater scale than the seasonal workers on which some farms rely – in many cases, migrants whose availability and desire to work can be dramatically affected by regional politics, conflicts, and socio-economic problems.

Automated harvesters of crops such as asparagus and mushrooms already exist, as do robots that can pick raspberries or strawberries from their plants in the field, while others can shake apples, oranges, and olives from trees. In some cases, autonomous or automated robots can focus on the best-quality produce, scanning each item for ripeness and flaws.

Autonomous robots can also help protect soft fruits from problems such as mildew, which can spoil popular, highly commercialised crops such as grapes and strawberries, which feed into other industries.

Specialist companies in these and other spaces include Dogtooth Technologies, Fieldwork Robotics, Saga Robotics, Engels Machines, and Ripe Robotics, with each company devising its own technologies. Some solutions allow fruit to be picked or treated autonomously at night, cutting down the need for expensive human labour in tiring, repetitive, but important tasks.

Land, crops, and livestock could be surveyed and monitored by ground sensors, helping farmers to gather data about how their holdings are performing in all seasons and weathers. Real-time, seasonal, and yearly data could help farmers irrigate the land, fertilise crops, and feed and water livestock in a more efficient, targeted, predictive, and sustainable way. More, it could help them plan their workforce needs in advance.

Farms could also be surveyed from the air by sensor- and camera-filled drones, operated autonomously or by remote pilots, helping to gather more real-time, seasonal, and comparative annual data.

Ground sensors and drones could then be linked to fleets of robots, and to smart, connected, and/or autonomous vehicles such as tractors, loaders, harvesters, and crop sprayers, sending the machines wherever they are needed most.

Linking integrated fleets of robots, drones, and connected vehicles to AI, data analysis tools, and Farm Management Software (FMS) could help create holistically managed, semi-autonomous farms in which farmers and on-the-ground teams become expert humans in the loop.

Over time, these technologies could also enable farmers build larger digital twins of their holdings, and thus help them manage feed and fertiliser needs in the longer term, improve crop yield and overall productivity, and anticipate both problems and opportunities.

Meanwhile, the journey from field to fork can be lengthy, complex, and environmentally costly when storing the harvest, then transporting it by road, water, or air. In these cases, other technologies, such as vertical farms and hydroponics could move some types of food production closer to the point of need in cities, where crops are grown indoors in ideal, sensor-monitored and AI-managed conditions.

In international export markets, distributed ledgers, blockchain, digital tokenisation, smart contracts, RFID tags, and temperature and humidity sensors could help authenticate, protect, and secure certain forms of produce, such as meat, extra virgin olive oils, and wine on long journeys from field to consumer.

Globally, the food chain is under pressure from a range of complex factors, including population growth, climate change, migration, political pressures, a worldwide drift from rural to urban living, plus an ageing population. In the UK alone, the number of citizens aged 65+ will increase from 12 to 17 million by 2035, while in the US, 65+ citizens will rise from 15 to 24% of the population by 2060. So, with global food demand expected to increase by seventy percent by 2050, sustainable farming practices are critical as production ramps up.

The flipside

All this is an attractive vision if it leads to higher-quality, healthier food being produced and shipped more sustainably to consumers, and to reduced waste, lower costs, a smaller carbon footprint, less reliance on pesticides, preservatives, and harmful chemicals, better yields and profits for farmers and growers, and lower prices for consumers.

This is the theory, at least, behind AgTech and increased farming automation, and many of these innovations are being implemented today, often in isolated or experimental deployments. But inevitably, the reality is as complex and nuanced as designing a robot to shake a tree, but without damaging or killing that tree.

Among the more obvious challenges, farming is slow, expensive, seasonal, labour intensive, deeply rooted in centuries of tradition and human cultures, and often lacks the capital to throw at high-risk, speculative technologies that might only be used periodically – for example, during the harvest season. More, such systems may be slow to develop and expensive to run and maintain, while offering no guarantee of long-term success or predictable returns.

While specially-designed warehouses and lights-out production lines are ideal environments for robots, cobots, and automated systems, farms by their nature tend to be less structured and less controlled, with uncertain terrain, challenging weather conditions, and all manner of physical obstacles to contend with. All this must be factored into robots’ design, safety, and performance, especially when working alongside people – not to mention vulnerable crops.

People are a challenge too: while the public is told that casual farm labour is menial, low-skilled work, in fact it is often highly skilled and specific: picking produce such as strawberries, mushrooms, or watercress demands knowledge of the produce and the correct techniques for picking it safely, cleanly, and efficiently. These are the types of skill that are hardest for any general-purpose robot to learn, and easiest to lose from the workforce through churn.

In many ways, therefore, automating farming is the hardest problem to crack with robotics, precisely because it is so hard to model and simulate consistently. In turn, that creates complexity at market level and, therefore, in financial and incentive terms. Would investors want to help commercialise a robot that picks one type of fruit in one country? If not, how would such a robot make it out of the university research lab that is working with farmers to solve real problems?

In the long term, this suggests that there could be viable applications for general-purpose human-equivalent robots: autonomous, dextrous, intelligent humanoids, which could carry out a range of different farming tasks.

But even this is not the only challenge. Farming is also prone to ‘black swan’ events: severe weather, climate change, blight, and the outbreak of viruses and infections, which could mean crop failure, the loss of livestock, and in extreme cases, famine. It is also highly dependent on macroeconomic factors, such as the price of fuel and the cost of other forms of energy.

None of this makes AgTech an ideal match for the kind of impatient capital that fuels the AI sector today, and which may undermine robotics’ ‘steady as we go’ progress as it leaches in from that sector. For years, robotics has been characterised by slow, iterative, standards- and safety-driven development in the pursuit of viable solutions to complex, real-world problems.

Faddy

So, while welcome in general terms – the sudden influx of venture capital to robotics and physical AI may put that at risk. Or it might distort the market by focusing investment on faddy, populist solutions rather than on where technology is urgently needed in human society.

Other areas are hidebound by regulation and safety standards – albeit for good reason. For example, flying Unmanned Aerial Vehicles or Systems Beyond Visual Line of Sight (BVLoS) is forbidden or restricted in many territories, while drones also need to be used safely and lawfully, and integrated into air traffic management systems.

But despite these caveats, the good news is there is much to celebrate in the AgTech sector, including some imaginative solutions to serious real-world problems. And robotics and AI are critical elements of these.

Professor Simon Pearson has an impressively long job title. He is Head of the School of Agri-Food Technology and Manufacturing, and Director of the Lincoln Institute for Agri-Food Technology, at the University of Lincoln – a city in Britain’s rural Midlands. As he explains, robotics, AI, and other forms of AgTech share a core function: environmental sustainability.

Focusing on robotics and AI at a Westminster Food and Nutrition Forum policy conference in December 2025, he says:

It’s very clear that the food system, particularly the UK food system, has all sorts of problems. Before Brexit, we had 63,000 seasonal migrant workers in the UK, but now we have licenses for just 43,000 and those numbers are falling. So, labour is an issue in the low-paid element of the job. But it’s an issue at all levels: in the middle-management layers and directors too, skills and the availability of labour are a big issue. We also have diet challenges: we need to get people eating more healthily – we've got an obesity problem in the United Kingdom. So, how do we do that? Well, the best way to do it is to make fruit and vegetables cheaper, so people buy more of them, and they are more available.

Automation, robotics, and AI are among the ways in which this could be achieved, he explains, by increasing yields and reducing waste.

Speaking in the same era as President Trump, who claimed last year that climate change is a hoax and a “con job” while the US AI Action Plan published in July 2025 talked of resisting “radical climate dogma”, Professor Pearson says:

We also have climate change problems. I spoke to a farmer a couple of years ago, and he said, ‘I've had six one-in-a hundred-year weather events in the last six years! Every year, there's so much climate challenge, so much uncertainty, and it's a very difficult environment for farmers to work in.

So, how do we solve those problems? We can't, says Pearson, but we can use technology to mitigate them.

Farming has always been a great adopter of innovation. It's a very, very adaptive sector. People move fast: farmers typically see a problem and look for solutions. They will adapt in any way they can, and often that is through technology.

But across the whole sector, one of the key things is productivity at all levels. That means economic productivity – reducing costs and increasing output – but also environmental productivity, reducing inputs into the environment. And it means social productivity, creating and protecting jobs in our industry, to have vibrant communities and vibrant companies.

So, how has productivity changed over the past 200 years, since the first Industrial Revolution? Sharing a chart on yields from 1800 to the present day, he says:

Up to the Second World War, we had an agricultural system which was primarily horse powered, and there was very little change in productivity. Then after the Second World War, several big events drove yield upwards. They included the availability of synthetic nitrogen [fertiliser] and synthetic pesticides.

On that point, the use of synthetic pesticides has become increasingly outmoded in recent years, because of the health risks to the public. Robotics could help counter that trend. In 2018, robotics and autonomous systems research group UK-RAS – part of the Engineering and Physical Sciences Research Council (EPSRC) – explained:

There is a global need to find new ways to produce crops that do not require, or which reduce the use of pesticides. There are now a number of crop-weeding robots that reduce the need for herbicides by deploying camera-guided hoes, precision sprayers, or lasers to manage weeds.Although in its infancy, this technology shows great promise. In addition, novel sensors deployed on robots can reduce pesticide use by both detecting pests and diseases, and precisely targeting the application of insecticides and fungicides. Robots could also be deployed as part of integrated pest management systems, for example, for the accurate and low-cost dispersal of bio-pesticides to counteract crop pests and diseases.

But the major productivity boon since the Second World War was the invention of the four-wheel drive tractor by John Deere in 1960, explains Pearson. From then on, there was more power from tractors than from horses. More good news, therefore. But there is a problem: while the global population is growing, productivity has been flatlining for several years.Pearson says:

We've hit the end of that phase of innovation, and we are seeing diminishing returns. But we've got to keep yield going up. It is currently about eight tons per hectare of land, but we could easily get to 14, 15, or 16. So there's a yield gap, and there is a great opportunity to keep growing productivity.

This is what AgTech should be delivering, the Professor believes. Robotics and AI are at the centre of these opportunities. For example, Saga Robotics – which trades as Thorvald, the name of its autonomous robot – spun out of the Norway University for Life Sciences and is now based at sites in Lincoln and Kent in the UK, in the US, as well as in Norway. Resembling white metal arches on wheels, fleets of Thorvald machines move along rows of strawberry plants or grapevines at night, applying UV light to destroy powdery mildew.

Meanwhile, a company called JABAS AI specialises in autonomy as a service: autonomous navigation systems for robots in farming and food production, especially for tasks such as spraying, weeding, and harvesting. Such technologies are necessary because in vineyards, in protective polytunnels, and even in open fields, traditional GPS navigation and mapping systems either don’t work or provide insufficient data about a robot’s location in complex terrain.

AI is at the centre of another AgTech innovation, this time from a specialist called FruitCast whose platform, backed by Ceres Agri-Tech, combines weather data, historical statistics, real-time crop images, and computer-vision-based analytics to deliver real-time yield forecasts for strawberry producers.  This is important for planning staffing levels and costs and, ultimately, for the availability and pricing of the produce itself. The company explains:

Strawberry growers face a constant balancing act: unpredictable weather, rapidly changing crop conditions and labour management pressures make it difficult to forecast yields with confidence. Too few pickers and fruit goes to waste; too many and costs spiral.

Today, the company forecasts yield for over 20 percent of the UK’s strawberry industry, so there is significant growth potential locally. However, its success also reveals that AgTech innovations tend to start locally and within specific niches. Even so, this technology could be developed for a broader range of crops.

ANELLO Photonics is a specialist company whose sensors and other tools optimise the precision of agricultural machinery, enabling farmers cut waste in resources such as water, fertilisers and pesticides, and pursue more efficient and environmentally responsible farming – even in GPS-denied environments. 

Meanwhile, Lincoln spin-out Agaricus Robotics has developed an integrated, dual-armed mushroom-harvesting robot, which can scan produce beds, select and pick mushrooms with minimal damage, then trim, weigh, and pack them.

As with all robots in these complex fields, they harvest data as well as crops: Agaricus provides crop analytics services, including forecasting, yield data, and waste reporting, helping to link automation with an efficient supply chain that grows smarter over time.

These are just a handful of the dozens of solutions out there in the market. At Lincoln, Professor Pearson explains:

Lots of people think, ‘Well, if I produce this widget, it'll solve all the problems in the world.’ But it never does. That’s because it's a systems challenge. It's about how you work with farmers, growers, entrepreneurs, and innovators to bring technologies to the market and get them adopted by farmers. The opportunity space in agriculture is immense. We've got robotics programs in lettuce harvesting, black grass, weeding, potato storage, robotic fruit picking, mushroom picking, and strawberry yield. But what I'm really trying to show is that the absolute need is huge.

My take

Time to dig in, perhaps, and invest for the future. But the need for patient capital is just as important as the need for innovation.