AI helps farmers turn data into better decisions

Artificial intelligence is helping farmers process routine data and could also support logistics, crop analysis and quality control, ERR reports. Researchers say the technology can speed up decisions, but farmers still need to assess its findings themselves.

Airika Verrev, crop and livestock manager and consultant at Pajo Ranch, said AI’s capabilities had improved significantly during the time she had used it. In spring, she tried using it to compile data in Excel, but it did not perform well enough. By harvest season, that had changed.

“I no longer make any Excel spreadsheets myself. I tell AI what I want and what it should look like, and there it is,” Verrev said.

She also uses AI to analyse data and search legislation for relevant provisions, but said the information it provides must be checked continuously.

Marten Madissoo, an associate professor of processing technology at the University of Life Sciences, said AI tools do not remove the need for farmers to judge the answers they receive. They can help people make decisions faster, he said, but users still need to be experts.

Madissoo sees significant potential for machine learning and AI across agriculture, including in logistics, the storage and processing of produce, and field analysis. He said there was no single solution for every farm, as the right approach depends on the problem being addressed.

Farmers already collect data for field records, registers and reports. Combining information gathered over several years and building a model to analyse it could help put that data to better use, Madissoo said. He stressed that agricultural and technical expertise must be brought together, with people who know the fields deciding which indicators matter and what they want to learn from the data.

From chatbots to field robots

Chatbots are only one possible application. Madissoo coordinates the work of AIRE, an AI and robotics centre at the University of Life Sciences focused on improving industrial companies’ competitiveness through the use of the technologies.

He cited an AIRE demonstration project at the Nõo meat-processing plant, where AI was tested for quality control of packaged products moving along a conveyor. The system was intended to detect problems such as a poorly attached label.