Ukraine’s largest broiler meat producer, MHP, said that it has developed an artificial intelligence (AI) solution to collect and analyse data from poultry houses to discover flows in the production process.
MHP said it took 5 years to develop its virtual assistant, which the company also calls “a virtual zootechnician”. The official name is Smart Technologist Assistant. “It’s like Siri, but in the poultry world,” the company explained. “A digital twin based on artificial intelligence helps employees in their work and warns about the risks of out-of-hours situations in poultry houses,” the company explained.
So far, the IT solution has proved its value in helping poultry farms to improve the uniformity of chickens in the flock, the accuracy of weight prediction and the mortality rate. However, as the assistant keeps learning, MHP managers are confident there is more to come.
“We connected deep learning algorithms and artificial intelligence and began to separate patterns, saw how poultry houses operate in general,” commented Nataliya Kondratenko, director of the MHP global IT expertise centre.
“We were interested in the technical indicators: whether our equipment runs correctly, how well the staff work, how to prevent a situation that can negatively affect the [production] indicators and quality of products,” she added.
One surprising thing the company discovered was that production performance depends not only on basic parameters such as climate control and feed quality. “There is also a concept of chicken happiness that should be taken into account,” the company said. The virtual zootechnician now has the crucial task of measuring the mood in poultry houses.
The system tries to identify what factors can cause concerns among poultry flocks in order to eliminate them, Kondratenko said. The key advantage of the new system, which primarily gathers data from video surveillance, is that it notices even small details that not every employee, even the most experienced, would typically pay attention to.
For example, the artificial intelligent solution managed to put together a correlation model linking changes in the chicken’s mood with the work of agricultural machinery or other equipment nearby.
“In general, some events sometimes seem unrelated [to production], but artificial intelligence says there is a correlation, and we can influence it. We did not think about some factors at all – that they can affect the quality of meat, but they do,” Kondratenko said.