AI Sorting Technology Targets Potato Post-Harvest Losses in Russia

Potato margins are notoriously thin, and several cost lines are moving the wrong direction at once. Freight costs have risen roughly 29% over the past year, based on figures cited from Forbes and ATI.SU, while EKSLi estimates that labor costs have nearly doubled over five years. Machinery, spare parts and other production inputs have climbed as well. An analysis published by Картофельный Союз (Russian Potato Union) frames this as a simple math problem: revenue has to grow faster than costs just to hold margins flat, in a business segment that was never especially high-margin to begin with.

Expanding planted area looks like an obvious response, but the piece notes it requires fresh capital and isn't always feasible — and a bigger harvest doesn't automatically mean bigger profit, since both product quality and market price weigh on the final number. That leaves post-harvest processing, and sorting in particular, as one of the few remaining levers an operation can pull without acquiring more land.

Potato sorting technology went largely unchanged for decades, with quality resting on the experience and attentiveness of the people working the line, and on how many of them were available. Buyers now expect more consistent quality, the cost of manual labor has risen, and seasonal staffing shortages have become a persistent problem — conditions the source material describes as a shared challenge across the industry rather than an issue specific to any one farm.

Computer vision systems are described as working the way a human sorter does, just at scale: cameras capture an image of each tuber, algorithms score it against defined parameters, and the system decides in real time whether to leave it in the main stream or divert it to a separate category. Accuracy is said to improve as the system accumulates more data. The source material also addresses a specific misconception directly — that optical AI sorting is still an unproven, "raw" technology — countering that such systems are already running in production facilities across Europe, North America and other markets with high quality demands. Russian growers are described as characteristically cautious about new equipment, generally waiting to see successful implementations at peer operations before committing capital, which is part of why live, on-product demonstrations are positioned as the most persuasive form of evidence.

EKSLi presents its offering not as a single machine but as a quality-management system spanning the full post-harvest chain, from intake after harvest through to the finished commercial lot. The product line includes:

• Agrosort H ("Урожай"/"Harvest"): primary sorting immediately after harvest, automatically identifying and removing soil clumps, stones, and green or rotten tubers; said to replace up to 12 workers on a line.
• Agrosort Q ("Качество"/"Quality"): used when forming the final marketable batch, flagging rot, mechanical damage, growth cracks, pest damage, shape defects and greening; said to replace up to 10 workers on a line.
• A secondary sorter called "Marzha+" ("Margin+"), used to recover second-grade potatoes out of the reject stream and to handle multi-stream sorting.
• Agroscan, described as an inline analyzer running on the conveyor itself.

Which defects get flagged depends on how the equipment is configured for a given operation's needs.

EKSLi's own estimate is that deploying computer vision and AI sorting can raise the share of marketable potatoes — volume that would otherwise be misclassified as reject — by a minimum of 5%. Around ten potato operations in Russia are currently reported to be running optical sorters in their processing lines, which the source material frames as the technology moving from early pilots into practical, everyday use. EKSLi also offers customers a calculator meant to estimate the total annual economic effect and payback period based on an operation's own cost and investment figures.

Albert Akhmetzyanov, an EKSLi expert quoted in the piece, ties the trend to buyer relationships rather than just production efficiency: "in recent years we've noticed that quality is no longer just a production issue. Today, it's the stability of quality that increasingly determines whether a farm can work with large buyers, meet their requirements, and build long-term partnerships."

Given that Russian buyers are described as preferring to see equipment proven on real product before investing, EKSLi has scheduled live demonstrations at three regional field days: Kostroma on July 30, Tver on August 5, and the Moscow region on August 7. At each event, EKSLi specialists are set to run optical sorters on-site, walk through current quality-control technology, and take questions on implementation from attending farm operators.