AI powered autonomous food safety management with anomaly detection

Companies mainly use HVT plans to assess the hazards, vulnerabilities and threats to their food production processes. For a medium sized company producing several products, there can be well over a hundred HVT plans that need to be created detailing all potential hazards/vulnerabilities/threats. However, companies generally use a combination of digital and inefficient paper-based systems. Therefore, company staff face a growing challenge to identify and manage all the risks before they become an issue. Coupled with the fact that the food industry has low profit margins, and that small and medium-sized businesses account for over 90% of the food manufacturing industry, companies lack the resources to manually manage this complexity. With product recalls in the UK for example costing tens of thousands of pounds, it is vital for companies to minimise the number of recalls by having accurate, up-to-date information on any risks.

To address these issues, the Consortium proposes to build on an existing food safety management software platform developed by lead partner Primority Ltd. This software acquires food safety and allergen alerts and uses machine learning to automatically extract key information from these alerts to detect potential outbreaks or large food safety incidents. However, this is not currently linked in with company HVT plans. We propose 2 key developments;

1. To develop machine learning algorithms that can automatically create HVT plans for different parts of a company's food production process
2. To develop anomaly detection algorithms that can link these automated HVT plans with Primority's proprietary analysis of food safety alerts and other food safety data sources to allow a proactive response to a risk and alert the company to any detected hazard/vulnerability/threat and the relevant HVT plan.