Conserve Italia: The Agronomic Network That Anticipates Risks in the Fruit Supply Chain

Conserve Italia is one of Europe’s largest agri-food cooperatives. Its fruit supply chain encompasses four crops—peaches, pears, apricots, and apples—grown by member farms located primarily in Emilia-Romagna, with some also in Puglia, Veneto, and Molise.

The Challenge: Analyzing a Fruit and Vegetable Supply Chain Through Data, Season After Season

Managing a fruit and vegetable supply chain means dealing every year with a variable that no company can control: the weather. Increasingly adverse weather conditions—late frosts, floods, and drought-stricken summers—affect yields, water consumption, and the severity of pests and diseases, with direct impacts on production and costs.

To address this variability, reliable and comparable agronomic data collected directly in the field were needed.

The goal shared with xFarm Technologies was clear: to equip the supply chain with an infrastructure capable of transforming what happens in the orchards into information that can inform decision-making.

The Solution: Three Streams in a Single Integrated Project

The collaboration between Conserve Italia and xFarm Technologies began in 2024 and is being carried out in partnership with the University of Bologna. Now in its third season, the initiative is structured around three complementary lines of work. 1. Digitization and monitoring network: an extensive network of weather stations and field sensors that feeds the xFarm platform with real-time agronomic data—currently one of the largest active agronomic monitoring networks in Italy. Each farm has been equipped with hardware and trained in the use of the platform through both online and in-person sessions.

2. Experimental agronomic project: On a group of pilot farms, strategies based on sensor data and models are compared with traditional management practices, measuring the actual effects on irrigation, pest control, and insect monitoring. 3. FarmView, fruit recognition and counting: a pilot project featuring a computer vision system—developed in collaboration with the University of Bologna—that counts fruits using images collected in the field and estimates yield throughout the growing season.

From the weekly bulletin to the risk calculated every 24 hours

During the 2025 season, the DSS models xFarm were systematically compared with the Emilia-Romagna Region’s bulletins: 6 models xFarm were compared with 16 regional bulletins and 79 zones, across four crops, from May through August. The key difference lies in temporal resolution: the regional bulletin is published once a week, and in the six days between one publication and the next, an infectious event can emerge, reach its peak, and begin to decline. The DSS xFarm updates the risk every 24 hours and, on average, signals an infectious event 4 days earlier than the first official announcement.

Where the comparison is methodologically sound—epidemiological models versus models—the agreement rate reaches 94% (15 agreements or partial agreements out of 16). And by identifying windows of low infectious pressure within periods classified as “high risk” by the regulations, a farmer guided by the DSS would have avoided 2–3 fungicide treatments per crop between May and July, resulting in potential savings on the cost of plant protection products for pear and peach trees and a potential reduction in the environmental impact. The purpose of the DSS is not to replace the regional bulletin, but to complement it: xFarm serves as a daily operational guide, while the bulletin serves as a regulatory reference for integrated production.

The Results: A Supply Chain That Measures Up

By the end of the 2025 season, the project will include 187 digitized farms, equipped with hardware and trained in the use of the platform; an extensive sensor network with 80 weather stations and 500 installed sensors; 4 crops monitored through experimental trials, with water savings measured compared to traditional management practices; and over 700 hectares monitored by the Agrometeo network. The FarmView pilot project made a significant leap forward between 2024 and 2025: fruit recognition accuracy increased from 56–66% to 76–81%, with approximately 40,000 images collected in the field, and a yield estimation error of 2.5% for pear trees and 4.3% for peach trees, calculated by integrating three years of historical data. For apricot trees, variability remains higher due to extreme weather events, while the 2026 roadmap calls for the integration of meteorological data and vegetation indices to strengthen the model.

Looking Ahead to the Upcoming Seasons

The analysis highlighted a principle that applies to the entire fruit and vegetable supply chain: given the same information, those who receive it first make better decisions. A four-day head start on an alert signal means more effective treatments, less waste, and decisions based on current data rather than snapshots from the previous week. The project’s third season aims to consolidate the results of the pilot, refine the yield estimation model by integrating meteorological data and vegetation indices, and strengthen supply chain reporting tools. The system developed by Conserve Italia in partnership with xFarm Technologies is not just a network of sensors: it is a digital infrastructure that collects data directly from orchards, assesses risks every 24 hours, compares agronomic strategies, and measures results over time.

The best decisions aren't made on the fly. They're carefully considered.