At IAOM MEA 2025 in Jeddah, millers, automation experts and grain analysts examined how AI is reshaping purchasing, quality, traceability and market risk and why data governance and human judgement still matter.

Artificial intelligence has moved rapidly from abstract concept to practical tool in the grain and flour value chain. At IAOM MEA 2025 in Jeddah, a panel moderated by Dan Basse, President of AgResource (USA), brought together voices from milling, automation and technology to examine how AI is already influencing grain trading and analysis and what comes next.

The panel included Ismail Jatti, Head of Manufacturing Excellence at Modern Mills Company (Saudi Arabia); Ali Magboul, CEO of ASM Process Automation (UAE); Sunil Maheshwari, Vice President Specialty at Siemer Milling Company (USA); and Anas Shaar, General Manager of EGI (Lebanon). 

The speakers agreed on one point: AI will not replace the miller, but mills that fail to modernise their data, automation and mindset risk being left behind.

FROM “SPREADSHEET HELL” TO LIVE DATA

All four panelists agreed that data is the foundation for any successful AI initiative and that many mills are starting from a very low base. Modern Mills’ Ismail Jatti described how, until recently, large parts of the decision-making process were driven by manual data collection and spreadsheet analysis. “We spent a lot of time collecting and analysing data just to make basic decisions,” he said. “Today, with AI and the tools available, we want to shorten that cycle: the data is there, the question is how quickly we turn it into decisions.” 

For automation specialist Ali Magboul, the main barrier is not technology but culture. Many flour mills still rely on Excel reports, disconnected lab readings and local spreadsheets. He calls it “spreadsheet hell”. His prescription is straightforward:

• Connect machine-level data from the automation and SCADA layers;

• Integrate it with the ERP system;

• Standardise and centralise it before introducing AI.

“ChatGPT is great in its world, but in a real industrial mill you need real-time, connected data,” he stressed. “AI will only be as good as the data pipeline you build.” 

DIGITAL MILLS READY FOR AI

On the operations side, Siemer Milling Company has already laid the groundwork. Vice President Specialty Sunil Maheshwari explained that their plants are now fully digitalised, generating millions of datapoints every minute. The next step is to bring AI into two critical areas:

1. Process optimisation and yield – using data to tune mill performance, improve extraction and stabilise ash.

2. Intake quality management – measuring protein, moisture, ash and impurities in real time as grain is unloaded from trucks, railcars or vessels, before it reaches the silos.

If AI can help mills segregate and blend grain more precisely at intake, he argued, it will deliver consistent flour quality and reduce the need for costly high-protein wheat. “We have been digitalising for three years; now the step is to let AI help us control the process,” he said. 

AI AS TRAINER AND SIMULATOR

For EGI’s General Manager Anas Shaar, the opportunity goes beyond analytics. His company has spent five years developing a dedicated AI agent for milling, trained on process data, technologies and real operating scenarios.

The agent plays two roles:

• Tutor: millers interact with it, asking questions and getting context-specific guidance. As they do, the system becomes more accurate and robust.

• Simulator: engineers can test new recipes, extraction targets or process settings virtually and see how they would affect flour quality before making changes on the plant.

“It has been a great journey,” Shaar said. “We are training the agent, and it is training our people.” 

The panelists emphasised that AI’s impact will extend far beyond the mill’s walls. Maheshwari pointed to traceability as a prime example. Legal requirements and consumer expectations are steadily moving towards full farm-to-fork visibility. Paper-based systems and fragmented spreadsheets can get nowhere near 100% traceability; AI-enabled systems integrating farm, logistics, storage and milling data can.

Moderator Dan Basse linked this to a deeper shift in consumer attitudes. Younger generations increasingly view “food as medicine”, demanding to know not only where their food comes from, but also its nutritional and functional qualities. With better integration between regenerative agriculture, farm data and milling analytics, AI could help connect soil health, wheat protein and amino acid profiles with the products that land in consumers’ baskets. “That is one of the most exciting possibilities of AI for me,” Basse said. 

FROM SOIL SENSORS TO PRE-HARVEST QUALITY FORECASTS

In practice, the first pieces of this farm-mill connection are already in place. Basse described how his company uses soil sensors and satellite imagery to estimate yields and likely quality – including indicators such as protein – before harvest. AI models can then aggregate and interpret this data, giving mills an early read on the coming crop.

That opens the door to pre-harvest contracting based on quality, not just volume:

• Sensors track fertilisation levels and soil moisture;

• Satellites assess crop vigour and disease stress;

• AI models predict yield and quality for each field.

Mills can then signal their needs – protein, gluten strength, functional parameters – and AI can help match them with farmers and traders, long before the grain hits the export pipeline. Bakers, in turn, will be able to use AI to refine water absorption, fermentation and baking curves based on the flour they receive. 

ROI: YIELD, QUALITY AND SMARTER BUYING

For the panel, AI’s business case rests on three pillars:

Yield and extraction: better process control means fewer losses and more saleable flour per tonne of wheat.

Quality consistency: more precise blending and process management reduce downgrades and customer complaints.

Smarter purchasing: early, data-driven buying can lower average raw material costs.

“Improving quality gives you a competitive edge; improving yield protects your margins,” said Modern Mills’ Jatti. “With these tools, I see a strong ROI. If we can buy earlier at competitive prices and run the mill more efficiently, the whole food chain benefits.” 

‘DATA IS THE NEW OIL’

One of the clearest themes of the session was that AI’s value in milling is inseparable from data governance. A senior CIO from First Mills stressed that industrial AI is a different proposition from consumer chat tools: it sits inside production environments, touches operational decisions, and therefore demands far stronger controls. To make AI usable at plant level, he argued, companies need three building blocks:

• Reliable process and automation fundamentals to generate consistent signals from the shop floor
• Data science capability to clean, structure and integrate data from multiple sources
• Robust cybersecurity and compliance, including local hosting requirements and national cybersecurity frameworks.

He also urged millers to treat data classification as a prerequisite, not an afterthought—clearly defining what is top secret, confidential, restricted or public—so proprietary know-how does not leak into open systems.

Basse echoed the point, calling proprietary data a company’s “moat” that protects long-term competitive advantage. Building on that, Ali Magboul captured the regional reality in one line that landed strongly with the audience: “Data is the new oil.”

REGULATION AND SYSTEMIC MARKET RISK

Beyond mill operations, AI in trading and markets raises systemic questions. Shaar warned that poorly regulated AI could have unintended consequences, citing examples from other sectors where people were wrongfully jailed based on AI-assisted decisions. In grain markets, AI-driven trading systems could amplify volatility or even destabilise markets if unleashed without oversight. He argued for diversification in AI models – using multiple agents and approaches rather than a single “black box” – and for clearer regulatory frameworks governing how AI can be used in trading.

CAN AI PREDICT GRAIN PRICES?

In response to a question from Miller Magazine about whether AI could one day forecast grain prices, Basse noted that models already exist which can keep prices within one or two standard deviations of fundamental “fair value” by combining crop, weather and demand data. But he also cautioned that many trading systems are momentum-based, leading to exaggerated moves as similar algorithms chase the same signals. “Machines can drive prices much higher or lower than we would expect from fundamentals alone,” he said, pointing to recent cryptocurrency swings as an example. Until better regulation and diversification emerge, millers should be prepared for more volatile markets in an AI-heavy environment. 

THE COST OF STANDING STILL

Perhaps the most human part of the discussion focused on resistance to change. Magboul told the story of clients who describe their automation systems as “too sensitive” to touch – a sign, he believes, of lack of awareness and unwillingness to learn. Jatti acknowledged that millers fear stepping out of their comfort zone, worried about problems they cannot yet foresee.

 Shaar argued that, in the end, competition is the strongest catalyst. “Resistance to change exists everywhere,” he said. “But when competitors adopt these tools and move ahead, the others realise they are falling behind. That creates real incentive to change.” 

WHAT AI MEANS FOR THE FUTURE OF MILLING 

For the milling sector, AI is not a distant abstraction but a practical set of tools that:

• Turn raw data into real-time, actionable insights;

• Enable pre-harvest quality forecasting and smarter wheat sourcing;

• Deliver tighter process control and more consistent product quality;

• Strengthen traceability and support the “food is medicine” narrative;

• Free millers from low-value tasks to focus on improvement and innovation;

• Force the industry to confront questions of data governance, privacy and regulation.