A recent smart manufacturing project involving Australian food manufacturer Kilcoy Global Foods (KGF), RMIT University and Food Agility CRC has delivered early results from applying digital twins and generative AI directly to food production. The project uses real-time production and equipment data to detect and assess anomalies, while providing frontline teams with relevant information and guidance to support decision-making.

Unlike more established applications such as machine vision, automated sorting and robotic handling, the project is not focused on automating a single production task. Instead, it combines digital twins with generative AI agents to create a real-time representation of the production environment, then uses equipment data, process information and operational knowledge to investigate and interpret anomalies.

The approach reflects a broader shift now taking place in food manufacturing.

For years, automation in food production has focused largely on repetitive and well-defined tasks. Packaging, conveying, palletising and inspection can all benefit directly from mechanical and automated systems, reducing manual work while improving throughput. As automation becomes more widespread, however, the challenge is increasingly moving beyond whether individual processes can be automated to how different machines, processes and data can work together more effectively.

At KGF’s Kilcoy facility in Queensland, a series of digital twins continuously monitor the production environment, while generative AI agents assess detected issues and determine which require human attention. Rather than sending every anomaly directly to operators, the system first captures and triages the issues, identifying those that require immediate intervention and those that can be reviewed later. This approach is designed to reduce alert fatigue while ensuring that critical issues reach the right people at the right time.

The project also demonstrates that the value of AI in manufacturing depends on more than simply collecting large amounts of data. The system combines operational data with ontologies, knowledge graphs and retrieval-augmented generation (RAG), allowing AI agents to interpret signals in the context of the process, product and plant rather than treating each data point in isolation.

Similar developments are gaining momentum across food and consumer goods manufacturing. In June 2026, Unilever announced plans to scale AI-enabled digital twins across its global manufacturing network, using live factory data and AI-enabled insights to identify issues earlier, improve quality and efficiency, and support faster decision-making. Industry coverage has also pointed to the growing use of digital twins in food manufacturing, extending beyond equipment visibility toward process optimisation, quality management and supply-chain traceability.

This suggests that intelligent manufacturing is entering a more practical phase.

The focus is gradually shifting from how much labour automation can replace to more fundamental questions: Can equipment generate useful data? Can that data be understood in context? Can different systems work together? And can the resulting insights support better decisions and more consistent production performance?

This shift may also change how food processing equipment is evaluated. Capacity, energy consumption, reliability and cost will remain fundamental considerations. But data acquisition, system interoperability, remote diagnostics and connectivity with production management systems are increasingly becoming part of the equation.

For food manufacturers, AI does not remove the importance of equipment, processes or frontline expertise. If anything, these foundations become even more important. AI can only deliver meaningful value when the underlying equipment, processes and operational data are sufficiently structured, reliable and well understood.

The more significant signal from projects like this is therefore not that AI will replace the food factory, but that digitalisation is moving deeper into the core of production.

From point automation to connected data, and from connected data to intelligent decision support, the next wave of productivity improvement in food manufacturing is extending beyond individual machines to the production system as a whole.