Over the last decade, consumers in the developed world have become accustomed to the next generation of smart devices. These are everyday objects such as televisions, watches, phones and houses that are interactive and connected through the internet. We now have the ‘‘internet of things’’. Smart technology enhances the user experience and allows manufacturers to add value to what are essentially everyday devices.

In the same time, we have seen an increase in the technology being used on farms. Indeed, the smart cow can now be added to the list of smart devices alongside smartphones, smart watches and smart TVs.

It could be said that the smart cow’s conception began 20 years ago with the introduction of feed-to-yield feeding systems. These use computer programmes to give more feed to cows that give more milk. More recently, automated heat and health detection aids have combined to grow the repertoire of decision support tools available to farmers.

Behaviour

These devices monitor cow behaviour and use computer-generated algorithms to predict the onset of heat, or if a cow is sick. The farmer can check, through his or her smartphone, the time each cow has spent eating, walking and chewing the cud. Sick cows and cows close to calving are identified and cows that are on-heat can be automatically drafted after milking for artificial insemination. In fact, a whole herd could be inseminated without the farmer looking at one cow for signs of heat or pressing a button on a drafting gate.

This removes the necessity of human decision-making in what is a critical part of the dairy farming calendar. But not only is the decision-making removed, so too is the physical labour – no tail painting, no time spent watching cows for signs of heat. These devices are also safer, as the period of artificial insemination can go on for longer removing the need for bulls on dairy farms.

Combine this technology with robotic milking and computerised calf feeding and the physical input required for dairy farming reduces even further. Research has shown that the actual time spent managing farms with high levels of automation does not change dramatically, but the type of work being done does, with much less physical work and much more data and information management type of work being done. This suits some people more.

Of course, the evolution of the smart cow hasn’t just been to do with removing the physical workload with dairy farming. Since 2009, there has been a quiet revolution in the way cows are bred. In the same way, geneticists can detect gene and health defects in humans using genomics, they can use genomics to fast-forward genetic gain in dairy cows by identifying bulls that have superior traits, from the DNA in a hair follicle. Over 70% of all dairy calves born in Ireland now have at least one parent that was genomically selected.

Future

Where to from here? If new technology in the last decade was about decision support tools and reducing physical labour input, then the next decade will be about artificial intelligence (AI). AI is essentially machine learning. In its simplest form, AI is giving a computer information (data) and making it come up with expected outcomes (results). By then feeding in the actual outcomes, the computer will get better at making predictions over time.

From a farming point of view, the possibilities with AI are endless. Let’s say every parish in Ireland had 10 fields with weather stations in them. These weather stations are just sensors monitoring rainfall and soil temperatures. This data, along with weather forecasts and grass growth information can all be fed into an AI computer. Grass growth doesn’t even need to be manually measured as the technology now exists to measure grass from space with the technology being used currently in New Zealand.

Using all of these parameters, the computer should be able to predict what grass growth will be like up to the next month. The amount of fertiliser being spread can also be fed in to the computer. GPS-enabled machines can tell the computer where it was spread and how much went out. Over time, the computer will learn from the outcomes and the accuracy of the predictions on grass growth will increase.

At the macro level, the implications of this are huge.

If growth is going to be slow, then feed manufacturers will know that extra feed will be required on farms so they can increase production in advance. Milk processors will know that if grass growth and ground conditions are going to be good, that extra milk deliveries will be coming. Not only will the volume of milk be able to be predicted, but also the fat and protein composition, as this is largely based on the diet of the cows.

From a supply chain management point of view, the implications of this are huge. Milk processing and the logistics around collection and end product deliveries will become much smarter. From the farmer’s point of view, feed and financial budgeting will become a lot easier and a lot more accurate.

That’s just one example. There are many more examples of where AI can have a huge effect such as predicting animal health problems, soil fertility, profitability, input requirements, environmental footprint and global supply and demand of dairy products. The key to all of this is sensors.

A cow on Liam Houlihan's farm, Bruree, Co Limerick, fitted with a heat and health detection sensor.

Next decade

Over the next decade, we will see the widespread rollout of sensors to farms. These will become cheap and almost disposable. They will be in fields, on tractors, in milking parlours and in cow’s rumens. Each sensor might only gather a very small amount of data, but they are connected through the internet to the AI computer that will use the data to compute outcomes.

Science Foundation Ireland last year announced a Teagasc-led project called Future Milk. The project is a collaboration between research and industry to develop future technologies for precision agriculture.

Over 50 research officers will be employed working on different projects across grass production, animal health, milk processing and genetics.

Effect

The question farmers will have is whether this new technology will improve profitability and for whom.

It is very easy to see the cost savings for milk processors and feed merchants as they will have visibility of what is coming down the line. This means they can plan for logistics, staff numbers, production levels and set targets for sales teams.

From the farmer’s point of view, the effect on the bottom line will be less visible. For sure, knowing grass growth and knowing future production has certain benefits but effectively this is going to happen anyway, whether the computer predicted it or not.

What the farmer needs to know is what to do and when to do it to get the best outcome – when to start grazing cows, how much meal to feed and when to plough the field and sow the crop. This will have real benefits for the farmer.

There are still a large number of unknowns. Who will pay for the sensors?

Who will pay for the AI and who owns the data? Is it the farmer generating the data, the tech company, or the milk processor?

The other concern farmers will have is about where the world’s food will be produced. Many developing and third-world countries have excellent resources in terms of soil type and weather, but lack the skills and human capital element.

Smart technology has the capacity to fast-forward human learning. Will technology enable a shift in food production away from the western world with its labour problems to the developing world in South America and southeast Asia where the cost of production is cheaper?