Aveleda pilots AI grape identification system in Penafiel

AI at winery intake

Aveleda is piloting Portugal’s first artificial intelligence system for identifying grapes at winery intake at its Penafiel facility. The project focuses on the point at which harvested grapes arrive at the winery and enter the intake process.

According to the available information, the system produces probability-based results rather than categorical decisions. The source does not specify exactly which characteristics or identification categories these probabilities cover. It also provides no details about the sensors, imaging methods, models or datasets used by the system.

Comparison with expert assessments

During the Penafiel trial, the AI system’s outputs will be compared with assessments made by experts. The evaluation is taking place under real harvest conditions, meaning the technology is being observed as part of an operational winery intake setting rather than solely in a controlled or laboratory environment.

No performance metrics, acceptance thresholds or trial duration have been disclosed in the supplied information. The number of samples, deliveries or grape types included in the pilot is also not stated, and the experts carrying out the comparative assessments are not identified. It is therefore too early to determine the system’s accuracy or how closely its results correspond with professional evaluations.

A pilot-stage project

The initiative is described as a pilot, and the available material does not indicate that the system has entered commercial deployment or broader operational use. No results, implementation timetable or possible next steps following the trial have been announced.

For now, the central purpose of the project is to examine how probability-based AI outputs compare with expert judgment during the practical conditions of harvest and winery intake. More information will be needed before the technology’s potential role in intake operations can be assessed, including data on its performance, consistency and the scope of the identification task.

The original information was published by Bing News – Vineyard AI.