AI-Driven Extraction and Systematization of Tacit Knowledge in Food Processing

CONTEXT AND PROBLEM

Tetra Pak, a global leader in the processing, packaging, and distribution of food products, is launching an Open Innovation challenge regarding the digitalization of Tacit Knowledge, which is the personal, experience-based "know-how" that employees have and that isn’t fully included or available on company manuals or databases.

The issue of Tacit Knowledge is increasingly relevant as expert technicians go into retirement, taking decades of situational "know-how" with them. For example, in the dairy industry, Tetra Pak provides food processing machinery and offers its clients detailed handbooks as well as on-site assistance and installation services. However, achieving a consistent high-quality product requires more than just following a standard recipe. Many variables can impact machine performance and the fine-tuning that ensures a prime product is often performed by technicians with years of experience.

Therefore, Tetra Pak is seeking to develop a sophisticated AI-driven system that can capture and analyze the knowledge currently stored in handwritten notes, local digital documents or emails in order to ensure that the technician's experience level doesn’t compromise the final product quality.

The solution sought must be GDPR-compliant, to guarantee that sensitive data is anonymized, as well as compliant with workers’ rights in the countries where Tetra Pak operates. It should also allow for interoperability between different tools, databases, and existing software. The solution should create a permanent digital legacy of Tacit Knowledge and mitigate the risks and costs associated with workforce turnover, including longer training cycles for new hires and potential increased production waste.

OBJECTIVES

• Scouting for a solution that leverages Artificial Intelligence (AI) and Natural Language Processing (NLP) to extract information from a variety of unstructured sources and that is able to distinguish between what is already present in the official documentation and what is Tacit Knowledge.
• Looking for solutions with TRL 3 and above
• Tetra Pak is interested in getting an overview of different technologies even if they can solve only part of the challenge as the solution sought is a system of systems (i.e. a set of complementary tools).