di Andrea Bergonzi, Data Scientist di Dataskills
Like all sectors, even the agri-food sector – among the largest and oldest in human history – is facing a phase of important change. The challenges it has to respond to are many, linked to the costs of production factors such as labour, raw materials and energy; the quality and traceability of the product; to its safety and wholesomeness and, increasingly, to its environmental, social and economic sustainability.
In a competitive context such as the current one, in a constantly changing global market and in a historical phase of very high awareness on the part of the final consumer, companies operating in the Food & Beverage industry are fully aware that Digital Transformation is the to follow in order to maintain or even improve one’s competitive edge and develop the resilience necessary to achieve business objectives.
And although the report “Agri-food 4.0: A survey of the supply chains and technologies for the future agriculture” published in 2020 still underlines a marked technological fragmentation that does not allow the full realization of an integrated Industry 4.0, it is now clear that between the technologies that most represent an opportunity to increase the quality of production, make it more sustainable and optimize costs, are those of automatic learning (Machine Learning).
In the Food & Beverage sector, the potential of Artificial Intelligence is enormous and companies are well aware of it: thanks to AI, it is in fact possible to automate processes and make the production chain more efficient in all its phases, from the raw material to the final distribution on the consumer table. In particular, this technology can support the important evolution process of the agri-food industry by optimizing four complex areas by nature: quality management, sustainability, decision-making processes, safety.
AI FOR AGRI-FOOD: HOW IT SUPPORTS QUALITY MANAGEMENT
Among the most onerous and complex operations for the Food & Beverage sector there is undoubtedly the management of product quality, understood as its correct classification and packaging. Furthermore, it is a repetitive process that would benefit enormously from automation and, not surprisingly, we are faced with a constant refinement of the neural networks applied to the sorting and packaging of food products: it is the so-called “detection”, and it uses precisely Artificial Intelligence.
A practical example of the application of AI in quality management in the agri-food sector is represented by the TOMRA systems, solutions based on optical sensors and HD video cameras that allow a significant reduction in the use of energy as well as waste, while ensuring the high quality of the operation. Artificial Intelligence also ensures a degree of accuracy even higher than 95% in the autonomous selection of products of different size, weight, color and shape, with greater product recovery compared to the use of traditional techniques (for example in the selection of potatoes starting from clods of soil and stones).
FOOD INTELLIGENCE AT THE SERVICE OF SUSTAINABILITY IN THE FOOD SECTOR
As already indicated, sustainability must be considered the guideline at the heart of innovation in all industrial sectors, and Food & Beverage makes no difference.
In this context, Artificial Intelligence can help reduce the environmental impact of the food production phase by collecting and analyzing Big Data from the entire supply chain.
There are many optimizations that AI can foster: the waste of resources, for example, is one of the main criticalities of the agri-food industry and must therefore be drastically reduced. In this sense, Artificial Intelligence can provide useful insights into the performance of breeding and agriculture activities – which can consequently be optimized following the identification of the so-called “avoidable waste”.
In crops, for example, Artificial Intelligence can exploit neural networks to analyze data from sensors that measure the parameters that influence production, such as solar radiation, humidity, temperature and the presence of chemicals in the soil. The cultivated soil will consequently be able to be better managed in planting, fertilizing and irrigation, ensuring a huge saving of resources. At the same time, the data generated by the AI can be used in a preventive and predictive perspective to maximize the yield of the land and the quality of the product.
AI TO SUPPORT DECISION-MAKING PROCESSES IN THE FOOD SECTOR
Even with regard to decision-making strategies – very complex in the agri-food sector due to the multiplicity of factors that can influence them – Artificial Intelligence can provide valid support. Timely and efficient decision-making is at the heart of the success of any enterprise, and its difficulty tends to increase the larger the size of the industry in which the organization operates. In this sense, AI offers business managers and administrators an analysis tool that is not rapid, but also extremely precise.
A practical example is the application of Fuzzy Logic (a Machine Learning method that simulates the functioning of human reasoning) in the coffee sector to optimize the bean roasting chain, a crucial phase of the supply chain for the quality of the finished product but influenced by a myriad of factors (the environment with all its variables, but also specific discriminants related to the grain itself such as shape, weight, size and humidity). The use of automation in this area has already made it possible to cut the costs of setting the correct processing parameters and, at the same time, also reduce its times.
And Coca-Cola has also used AI to support its decision-making processes, using Deep Learning to analyze the data collected by self-service machines that served drinks that the consumer could customize to his liking by combining different flavours. It goes without saying that the analysis of these Big Data has proved to be crucial for improving existing products as well as for creating new ones (an example is the Sprite Cherry, still little known in our country).
SAFETY IN THE AGRI-FOOD SECTOR AND ARTIFICIAL INTELLIGENCE
Finally, Artificial Intelligence can provide its fundamental contribution in the field of product safety, which is essential for obvious reasons for any company operating in the agri-food sector. Never as in this historical phase is it in fact important that foods comply with increasingly stringent sector regulations as well as health safety procedures along the entire production chain, which must not undergo any contamination in order not to endanger the health of the consumer .
In this context, AI can act on several fronts: detecting the correct use of PPE by employees through real-time monitoring; ensuring the effective (and faster!) sanitization of machinery and equipment with equipment such as optical vision systems, sensors, ultrasound and AI-powered lasers and with Machine Learning systems that process the information collected to determine the quantity and quality of debris food and microbial substances left in the machines.
TECHNOLOGICAL INNOVATION IN THE FOOD SECTOR: WHERE IS ITALY?
Having clarified the numerous aspects in which Artificial Intelligence and, more broadly, technological innovation can support the evolution of the agri-food sector, we discover that Italy still has ample room for improvement.
In fact, in January 2023 the report “Investments in agrifood-tech in Italy 2022” was published, prepared by TheFoodCons in collaboration with Agrifood-Tech Italia, which highlights a not exactly rosy situation. Italian investments in the sector of new technologies in the food sector exceeded 156 million euros in 2022: a figure that might seem impressive, but which in fact represents only 0.30% of world investments in the sector, equal to approximately 52 billion dollars (of which 10 billion euros in Europe alone).
Furthermore, these resources were mainly dedicated (about 41% of the total) to the last phase of the food supply chain, i.e. D2C consumption models (from producer to consumer), E-Grocery, Marketplace and Delivery. On the other hand, 38.9% of resources were invested in Agritech, particularly in the areas of Farm Management, Novel Farming and Indoor Farming. Finally, in third position is the technological area dedicated to Ho.Re.Ca., with investments equal to 11.3% of the total.
At present, our country – which is in third place in Europe for gross salable food production – still needs a more agile and collaborative approach to the sector: this is the only way forward for the Italian agri-food industry to truly become resilient and can successfully face future challenges.
ARTIFICIAL INTELLIGENCE AT THE SERVICE OF FOOD & BEVERAGE: SOME VIRTUOUS EXAMPLES FROM AROUND THE WORLD
We conclude this excursus with a short list of virtuous examples of the application of Artificial Intelligence in the agri-food sector from all over the world, to highlight in which and how many ways this technology can contribute to the innovation of a sector that is not in a position to stop to evolve.
Valio: This Finnish dairy company used AI (which analyzed data from global chocolate-lovers) to design a low-sugar milk chocolate.
Rio Bravo Brewing Company: is a New Mexico brewing company that used ChatGPT to formulate the recipe for its new beer, Alegorithm.
Sapporo Holdings: In collaboration with IBM Japan, created an AI system that creates pre-packaged cocktails starting from the generation of one hundred possible formulations. Recipe generation takes place in seconds.
Lunchbox: is a startup specializing in food technology that uses Food Intelligence to improve restaurant menus, increasing the sales of its customers.
Nestlé: the Food & Beverage giant has long been making significant investments in Artificial Intelligence and Machine Learning, in particular to improve its R&D process. From 2016 to today, it has recorded a process acceleration of 60%.
The ones just reported are just some of the numerous examples of implementation that this technology can bring to a sector that is not made up only of the mere production of food and beverages, but of a complex ecosystem that also includes marketing, purchasing, quality control, optimal supply chain, packaging, logistics, transport.
A trend destined to increase in the coming years as well, especially in consideration of the speed with which Artificial Intelligence systems evolve.
