In highly regulated, traceable supply chains like those enabled by KPMG Origins, there are few entry points for inappropriate material or conduct. One major risk however occurs during “generation/production” stages —such as the growth and harvest of crop in agriculture or production in manufacturing — where creation or conversion rates are often poorly understood, even by experts, due to specific context and varying conditions. For instance, in an agricultural example, crops from deforested land might be sent to a compliant farm, which then falsely claims them as its own. To combat this, KPMG employs up to 9,000 auditors in Brazil, though monitoring vast, often compliant, remote farms remains a challenge. A growth model, factoring in farm and crop-specific contexts, can help by identifying suspect yields in real time and prioritising audits. These models were trained on public and discernible information (e.g. satellite feeds) data points, against TDM counts from auditors. This project advanced to a real-world proof of concept, integrating KPMG’s auditing expertise to validate the model’s effectiveness in enhancing regulatory oversight and traceability.