Wheat is one of the most widely used ingredients globally. Yet, nature is inherently inconsistent: grain quality varies by field, by season, and even within the same shipment. For millers and bakers, this inconsistency presents one of the most challenging tasks: how to deliver flour that behaves predictably in every bakery, every day.

For centuries, the industry has invested heavily in testing and measurement to understand wheat functionality, track protein levels better, classify wheat types, and adjust with additives or blends. Yet most of these efforts take place only after the grain has already reached the mill, when variability has already entered the system. Millers dedicate significant resources and expertise to managing it, but despite their best efforts, they lack the technology needed to predict flour behavior with accuracy. In an industry thousands of years old, the question is straightforward:

Can artificial intelligence bring new precision to an age-old challenge? 

Stybel’s Challenge

Stybel, a leading milling company in Israel, operates five facilities nationwide. Up to 90% percent of its wheat is imported, which means variability is unavoidable. Even with modern milling facilities, including its state-of-the-art Milling site in Ad Halom, the company strives consistently to translate grain quality into predictable flour functionality. 

Flour functionality, how dough absorbs water, how long it mixes, how stable it remains, and the final texture of baked goods, is what ultimately matters to customers. However, as Omer Thon, a fourth-generation miller and part of Stybel’s management team, explains, protein alone cannot predict these outcomes. The company needed a way to bridge the gap between grain specs and bakery performance. 

“Consistency is the core of what our customers expect from us,” says Thon. “If the flour behaves differently every week, bakeries can’t operate efficiently. That is the problem we set out to solve.”

Enter Equinom

Equinom, a 14-year-old food-tech company specializing in plant protein biochemistry, saw an opportunity to solve this long-standing problem. Equinom developed Manna™ for Wheat, an AI-powered platform designed to predict wheat functionality with accuracy and speed.

The platform operates by combining Near-Infrared (NIR) spectroscopy, a technology widely used in the grain industry, with Equinom’s proprietary machine learning models. Instead of focusing on a handful of specs, such as protein or moisture, Manna™ captures the entire spectral fingerprint of each wheat lot, translating thousands of data points into predictions about real-world performance.

By integrating spectral data with historical milling and baking records, the system generates actionable insights in seconds. This enables millers to make more informed decisions regarding purchasing, binning, and blending, thereby avoiding surprises in production.

Changing the way a mill works

Stybel implemented Manna™ as part of its daily operations. NIR devices scan trucks arriving at the mill, and the Manna™ platform immediately predicts the grain’s quality. Instead of storing wheat by origin, Stybel now stores it by predicted functionality, directing each load to the most suitable silo.

This shift‚ transformed bin management. Bins that were once underutilized now operate at higher capacity, while grain is allocated more flexibly. As a result, Stybel has improved both its storage efficiency and its product consistency. 

“In the past, we worked around variability,” Thon explains. “With Manna, we don’t just manage variability, we use it to our advantage. We see the strengths of each lot and put them to work in the right place.” 

The platform also supports blending. By analyzing flour specifications and functional parameters such as dough rheology, Manna™ suggests optimized blends that meet quality requirements at the lowest cost. This not only reduces reliance on trial-and-error blending but also ensures consistent flour performance for customers. 

Implementation and Ease of Use

Introducing AI into a traditional industry is not without challenges. The most significant hurdle, as both Stybel and Equinom note, is changing the mindset of an industry accustomed to doing things a certain way. Yet, the process itself is straightforward. 

Before installation, Equinom builds a customized model for the milling site, using the customer’s historical data. The model is then validated and refined with real-world feedback. Once in place, the system requires no additional hardware beyond existing NIR devices, making adoption practical and cost-efficient. 

Beyond Milling: Supporting Bakers Too 

Equinom works globally to support millers and traders in improving product consistency and reducing costs. Recently, the company expanded into the baking industry, where the cost of inconsistency is felt most acutely. Bakers suffer from unpredictable flour performance, even when flour meets specifications. Manna™ now helps bakeries increase yield and reduce waste by ensuring flour behaves as expected. 

As bakeries embrace more innovative tools to protect their margins and meet consumer demand, the bar for consistency is being raised across the supply chain. For millers, this new reality underscores the importance of delivering flour that performs reliably, every time.

The Bigger Picture: Beyond Protein Pricing

Across the milling industry, there is broad agreement: protein is a limited predictor of flour quality. Still, global wheat pricing is based almost entirely on protein levels. This mismatch leaves significant inefficiency and cost across the supply chain.

Equinom’s vision is ambitious but clear: to replace protein-based pricing with functionality-based evaluation. By equipping millers, traders, and bakers with the ability to assess quality in seconds, Manna™ could shift the economics of the entire wheat industry. 

As Thon reflects, the potential is transformative: “If the supply chain could buy and sell wheat based on real quality, not protein, it would change everything, from how we purchase vessels to how bakers plan their production. It’s a win for everyone.” 

A New Era of Milling

For Stybel, adopting Manna™ was more than a technology upgrade; it was a strategic leap. By moving from reactive adjustments to predictive management, the company has secured its reputation for consistency while embracing innovation. This success story shows that even in one of the world’s oldest industries, data and AI can drive meaningful transformation. What began with one mill in Israel could become the standard for millers everywhere: a new way to turn nature’s variability into predictable, high-quality flour.

‍