By Dan Jacobs, Senior Editor, AgriBusiness Global and CropLife
The way artificial intelligence has exploded into popular culture is, for some, equal parts exciting and terrifying. For agriculture, the fact that this new technology promises to solve a variety of problems, including labor, has many industry experts leaning to the side of excitement.
“Artificial intelligence is already making significant strides in mitigating labor shortages across various sectors of the agricultural supply chain…"
In just about every survey that asks about the ag supply chain’s biggest challenges, finding enough labor is a constant and universal issue. While it might not completely solve that problem, artificial intelligence (AI) has the ability to help.
“Artificial intelligence is already making significant strides in mitigating labor shortages across various sectors of the agricultural supply chain,” says S. Daniel, Founder at Vellore, Tamil Nadu, India-based Grow Iota — an AI-based company with the motto of “AI is for Everyone.”
“In manufacturing, AI-driven automation is streamlining production processes, reducing the reliance on manual labor. In distribution, AI-powered logistics and inventory management systems help optimize routes, minimize downtime, and improve efficiency. At the crop input end, AI is enhancing precision agriculture by optimizing resource allocation and reducing the need for extensive manual intervention,” he says.
Agriculture has certainly embraced technology in many forms. But AI’s ability to learn as it grows is still a new concept for everyone (in every industry), and that means companies are still working out how it can be used. For Alina Piddubna, AgriTech Portfolio Manager with Intellias, AI will play different roles for different tasks.
“Depending on the nature of work to be done, the approach can be different, and the impact can be different,” she says. “Of course, artificial intelligence will help reduce the operational costs.”
Nitin Bhatia, Senior Partner at Headstorm, agrees that AI has already found its way into some companies focused on agriculture.
…the system examines the multiple variables and the results produced and compares them to the desired outcome. The more data the system internalizes, the closer it will get to those desired results.
“Artificial intelligence is a very, very broad discipline,” he says. “And under artificial intelligence there are a whole slew of technologies. You have heard about machine learning, computer vision, and natural language processing deep learning. But what’s important to understand is, how do you apply those technologies? Some are fairly straightforward, so, some companies [for example] are already starting to use some type of predictive analytics to access their specific situation and act accordingly.”
The Promise of AI
AI has burst onto the scene, with a rapidity and promise that suggests the agriculture industry will see an instant transformation. However, while there has been some integration and change, sea change is further away.
“We have this great vision of the future, and we think it’ll happen tomorrow,” says Paul Streater, Chief Operating Officer, AgBioScout, a London-based intelligence and advisory company. “But the truth is, it’s going to take probably several decades to happen.”
Companies are still figuring out how AI fits into their operations. What separates AI from a simple software solution is its ability to continually learn from previous outcomes.
“We’re seeing AI start to penetrate in propositions coming in the market where they’re pulling a lot of data points,” Streater says. “And the great thing about AI is you can handle it (better) than a human brain does. It can handle a lot of data very quickly.
“[And it can] interpret that. And it can learn from an interpretation. That’s different,” Streater continues. “A true AI probably will learn from the data.”
In other words, the system examines the multiple variables and the results produced and compares them to the desired outcome. The more data the system internalizes, the closer it will get to those desired results.
Streater shared the results of one manufacturer’s test of AI pitting human agronomists against an AI.
“The outcome was [the company’s] agronomists got the answer right 75% of the time whereas the AI was in around 90%, Streater says. “So, you’d never have thought that — for a traditional business, like farming, there’s no way a computer’s going to outperform our highly trained agronomists. That just goes to show that technology is evolving. And it can start to really make a difference in the detection of issues with crops, but also understanding a lot more data points as well and learning.”
AI has already made its way throughout the supply chain, though execution is still in its infancy.
Streater doesn’t expect the role of human agronomist to go away. AI will take in the metrics and offer recommendations.
“I think it will always have one human there to oversee and see if that makes sense. I trust there’ll always be some human element. But the majority of that crunching work will be done by AI.”
Along the Supply Chain
“Many companies are actively exploring AI applications to enhance supply chain management,” Daniel says. “Predictive analytics, demand forecasting, and real-time monitoring are some examples. AI enables proactive decision-making, reducing delays and ensuring the smooth flow of goods from farm to table.”
AI has already made its way throughout the supply chain, though execution is still in its infancy. Companies are exploring how this new technology might solve their issues, including labor challenges.
“Agriculture companies are in different stages of adoption of artificial intelligence,” Intellias’ Piddubna says. “Every time you consider implementing the solution, you need to be aware of which operational processes and algorithms must be adapted to harvest the most value of it.”
Bhatia sees progress coming in stages. “Companies are starting to think, ‘Okay, how do we leverage drone technology, robotics technology,’” he says. “And depending on whether you’re talking broad acre or specialty, that technology will vary. The way you solve that problem will also vary. That’s how I think companies are going to continue to see the small waves of technology come in and solve those problems in layers.”
One reason labor is an issue, at least at the farm level, is the aging population. Fewer young people are interested in following their parents into the business. Not only can technology help mitigate that lack of labor, it might also have a secondary effect, Piddubna says.
“Artificial intelligence increases the efficiency of the process of production, not only in terms of reducing the operational cost, but it also increases the attractiveness of the of the industry,” she says. “The more advanced technology you have, the more people you can attract who are interested in the application of this technology.”
For that to happen a number of entities from governments to universities must adapt their approach.
Changing the Investment Spend
“The integration of AI in manufacturing is changing the landscape of investments,” Daniel says. “Companies are reallocating resources toward AI-driven technologies, robotics, and smart manufacturing processes. These investments are aimed at improving productivity, reducing labor costs, and increasing overall competitiveness.”
Spending on AI in agriculture is projected to grow from $1.7 billion in 2023 to $4.7 billion in 2028, with a CAGR of 23.1%, according to Research and Markets. Ag-related supply chain companies working within agriculture should dedicate some part of budgeting to that, Piddubna suggests.
While AI is a hot topic, the cost of investing will be prohibitive for some companies.
Ultimately, AgBioScout’s Streater expects to see a three-tiered system for the use of AI in agriculture.
Tier one includes wealthier countries that can afford the investment producing high-value crops. “I could see that replacing a lot of labor shortages,” he says.
Tier two includes wealthier countries, but smaller operations like those found in many parts of Europe where “we might have a blend of human and robots,” Streater continues.
The final tier includes smaller farms in poorer countries where so much of the population relies on farming. “It actually supports the whole economy. You cannot see robots particularly penetrating that area, and labor is relatively cheap. So, it’s going to be a balance between that labor cost and the farms where they’re operating,” he says.
“A smart company will go cautiously into that, not putting all the [money] into one nest,” Piddubna says. “A company would start from a pilot [project], trying to check the application of the selected technology solution with artificial intelligence, which is embedded into some point across the supply chain to prove their result. That’s what we recommend to our clients to start from.”
Mitigating labor problems is one benefit, but companies must be careful to balance the potential benefits without eliminating the human element. And that is one place government regulation could help.
If there is a problem with artificial intelligence it’s that developers of the technology don’t understand the need to “balance between planet, people, and profit.".
“The impact will be constrained partially by government regulations, especially by those dealing with social protection,” Piddubna says. “So, for example, if you’re speaking about France, I’m sure that the Ministry of Labor, Employment and Economic Inclusion would always interfere trying to balance this impact so they will not allow huge adoption (of AI) in an uncontrolled manner. Companies across the supply chain adopting AI technology will always be balanced by the labor regulation policy.”
Challenges
If there is a problem with artificial intelligence it’s that developers of the technology don’t understand the need to “balance between planet, people, and profit,” Streater says. “We want to make things more sustainable, but we can’t actually take away rural livelihoods or take away profit because you need profit for innovation, to keep business going. So, there’s this fine balance between those two worlds — the traditional way of farming and industrialized farming model. And there are those trying to disrupt it. They need to work together as it’s the same problem for everyone. We need food on the table.”
“It’s essential to consider the potential challenges associated with AI adoption, such as data privacy, cybersecurity, and the need for workforce upskilling,” Daniel says.
Daniel cites another concern often raised around technology.
As advanced as the technology may be, there are some basic problems that must still be overcome, Bhatia says.
“The machines are getting smarter and smarter and smarter,” he says. “From a manufacturing perspective, they’re trying to embed as much of that intelligence on the edge, meaning the machine can make the decisions for you in real time versus having to ship that data back and forth. That comes down to connectivity. A lot of the farmland in the United States does not have great connectivity and that can tend to become a bottleneck.”
With any new technology, the expectation and the reality don’t always run parallel and often unforeseen roadblocks appear that must be surmounted.
In other words, it will likely take large financial commitments at both the private and public level before many of AI’s labor-saving solutions are realized.
“If you speak about the impact on labor, it reduces the operational costs in the long term, not in the short term,” Piddubna says. “So, we need to understand that it’s an investment which must be cautiously decided upon. But it is time to start now, to generate benefits for the future.”
With any new technology, the expectation and the reality don’t always run parallel and often unforeseen roadblocks appear that must be surmounted. Bhatia cites another relatively new technology as an example.
“Once cloud [storage technology] came into the picture, we could scale a lot of these different solutions,” he says. “We could capture as much data, however, as we wanted to. So, a lot of companies have gotten to the point where they have all these solutions deployed into the cloud fully scalable, fully extendable. They’re capturing all the data. But now what’s happening is all this data is sitting in silos.
“What needs to happen next is starting to mesh these solutions together — create a common data language and common data platforms,” he continues. “Once you create a common data language and common data platforms on top of that is where you can start to build a lot of the machine-learning models, the deep-learning models, the intelligence that you need to provide the farmer the information. There are lots of gaps in the data, whether it’s from a coverage perspective or activities perspective. And that’s why the accuracy of those predictive analytics is not that high, but as we bring the data together and fill in some of those gaps, I think the accuracy of those predictive models will get a lot better.”
