SORTEX AI700

SORTEX AI700

Seeing the difference: AI redefines optical sorting 

Sorting oats for gluten-free markets is one of the most demanding challenges in food processing. When defects look almost identical to good grain, traditional color sorting reaches its limits. The new SORTEX AI700 is Bühler's most advanced optical sorter yet, a deep learning machine that moves beyond color and shape to see what no sorter has seen before. From greater food safety for consumers, to improved profits and reliability, it is a gamechanger for food processors. 

In food processing, not all defects are easy to see. Some look almost identical to the product they contaminate, especially in grains such as oats, barley, and rye. For processors supplying gluten-free markets, visual similarity creates a serious safety and commercial challenge. This is where artificial intelligence (AI) can be a gamechanger. 

Traditional optical sorters, which rely primarily on color and shape, perform well when defects are visually distinct, but are pushed to their limits when sorting visually similar product. Bühler’s new SORTEX AI700 is a leading example of AI usage in food processing that helps to bring safer food to consumers, and better yield and profitability to millers. 

The SORTEX AI700 is an optical sorter that doesn’t use chatbot AI, but instead dives into the world of Convolutional Neural Networks (CNN) and deep learning. With these tools it analyzes images at pixel level, learning to recognize subtle patterns that distinguish acceptable product from defects, even when they look almost the same. 

With this technology, SORTEX has been able to achieve the highest yield quality ever in oats, and the journey into other products is just getting started. 

Why change is needed 

In recent years, the pressure from food standards has intensified. Processors are contending simultaneously with rising input costs, tightening supply chains, escalating regulatory demands, and a consumer base that is better informed, and more demanding, than at any point in history. The industry is being asked to do more, with less margin, to a higher standard. Technology is not optional in that environment. It is the difference between a business that adapts and one that does not. 

Beyond the heightened consumer and regulation standards, climate change is altering harvest conditions across every major growing region. Prolonged droughts, flooding events, and temperature extremes are increasing the prevalence of damaged, diseased, and contaminated grain in ways that were not anticipated when many existing processing lines were designed. Defect profiles that were once predictable are becoming less so. For processors relying on sorting technology configured for a stable input, that unpredictability is a growing liability. The equipment of yesterday is being asked to solve the problems of tomorrow. 


SORTEX AI700 provides real-time data on raw material quality, empowering decision-making. 

How AI is changing optical sorting 

Melvyn Penna, Senior Product Manager at Bühler SORTEX.

Place a grain of oat, barley, and rye side by side, and the human eye struggles to tell them apart. Now tip a kilogram combination of the three onto a conveyor belt, and the challenge of sorting them becomes all but impossible. This is the opportunity of AI, and the inherent limitation of traditional color sorting. 

The SORTEX AI700 breaks through this limitation. At its core is a CNN, the same class of AI that underpins image recognition in advanced medical diagnostics and autonomous systems. Where a traditional sorter asks, “Is this grain too dark?”, the CNN processes color, shape, and texture together, building a comprehensive picture of every individual grain on the belt. 

“In the past, our machine learning systems relied on engineer setup and tracking algorithms to improve upon initial parameters,” says Melvyn Penna, Senior Product Manager at Bühler SORTEX, an expert who has spent 35 years at the forefront of optical sorting innovation. For Penna, the AI700 is the most exciting project yet. “Now, with the AI700, the machine is trained with millions of labeled images. It develops a deep understanding of what is acceptable and what is not. It doesn’t just see a defect; it knows a defect.” 

The gluten-free challenge 

The AI700's first application targets one of the industry's most persistent problems: removing gluten-containing grains from oats. This is an area where safety meets commercial pressure.  

Oats are naturally gluten-free, but they are routinely co-harvested and co-transported with barley, wheat, and rye, all of which are allergenic. Visual separation is extremely difficult. The consequences of failure - product recall, regulatory sanction, and harm to the consumer, are severe. 

“We’ve achieved unprecedented levels of removal for barley, wheat, and rye from oats,” says Penna. “Traditional sorters, while meeting regulatory standards, cannot match the precision of the AI700. For processors supplying certified gluten-free markets, that gap matters enormously.” 

It matters commercially, too. Higher output purity means more product reaching the market, stronger relationships with retailers and certification bodies, and reduced write-offs from rejected batches. The AI700 achieves this not by rejecting more but by rejecting more accurately. Its CNN-powered classification resolves uncertainty with understanding rather than caution, accepting what should be accepted and rejecting only what should be rejected. 

The numbers make the case clearly. “A 5-10% improvement in reject concentration, sustained over a full production year, can translate to hundreds of thousands in revenue that would otherwise have been lost,” says Penna. “The yield lost or gained is a key catalyst in the success of a business.” 

For processors in the gluten-free segment, this is not a marginal gain. It is a structural improvement in how the line performs and in the quality assurance story they can tell every partner in their supply chain. 

Built to work from the start 

While traditional sorters require specialist setup and ongoing calibration, the SORTEX AI700 is designed to eliminate that dependency. 

Because it arrives pre-trained on millions of images, minimal manual configuration is required. The machine is ready to sort accurately from day one, applying consistent, deep-learning-derived decisions regardless of who operates it or when. Shift changes, staff turnover, varying operator experience, none of this affects output quality. 

“It essentially works straight out of the box,” says Penna. “The risk of human error is removed. Consistent quality is guaranteed regardless of shift or operator experience.” 

In a food safety context, that consistency is critical. A sorter that performs differently depending on how it was set up introduces variability into a process that demands certainty. The AI700's pre-trained intelligence eliminates that variability at source. 

Shaping the industry’s future 

The AI700 is not a response to a single customer need, it is a response to a convergence of pressures that are reshaping the economics of food processing. 

Raw material costs are rising. Every percentage point of improved yield carries greater financial weight than it did five years ago, while climate change is increasing the variability and prevalence of defects in harvested grains.  

“The future presents global challenges in food processing which, if left unaddressed, will damage both processors and consumers,” says Janine Wegmann, Head of Product Management at Bühler SORTEX. “We built this machine to answer the questions processors are already facing, and the ones they will face tomorrow.” 

The AI700’s architecture is built for this uncertainty. Trained initially for oats, it is already being prepared for expansion into almonds, other tree nuts, and additional grain applications. Each new commodity capability will be deployed via the cloud, meaning that machines in the field today will become more capable over time, adapting to new markets, new regulations, and new challenges without the need for replacement. 


Janine Wegmann
We built this machine to answer the questions processors are already facing, and the ones they will face tomorrow.

Janine Wegmann,
Head of Product Management at Bühler SORTEX

Already delivering results 

The SORTEX AI700 launched in London in May 2025 and is already operating in oat processing facilities. Early adopters in the gluten-free market report improved defect detection, yield, and ease-of-use. 

“This is the most sophisticated optical sorting solution we’ve ever developed,” says Wegmann. “SORTEX AI700 doesn’t just meet existing standards, it sets new ones.” 

For processors ready to move beyond the limitations of color sorting, AI700 offers a clear and proven path forward: safer product, better yields, smarter operations, and a platform that grows with the demands of the industry. 


contact block

How can we help?


Gupfenstrasse 5
Uzwil
9240
Switzerland