Recipes supported by AI

The SmartBakeChain (SBC) project of NEURA GmbH, based in Bremen, pursues a clear objective: to systematically digitalize artisanal baking knowledge and make it usable through artificial intelligence. Against the backdrop of fluctuating raw material qualities, increasing quality requirements, and a growing shortage of skilled labor, SBC aims to establish a new, scientifically grounded foundation for stable and reproducible baking processes.

Initial Situation and Challenges

The baking sector is characterized by heterogeneous and often isolated digitalization approaches. Conventional software solutions primarily focus on recipe management, production planning, or documentation. However, the complex interactions between raw materials, process parameters, and product quality are rarely modeled. As a result, essential experiential knowledge remains implicit, leading to inefficiencies in process control, quality assurance, and innovation capability.

Core of the SBC Project

SmartBakeChain directly addresses this gap. Within a feasibility study, a modular, AI-based quality model for baking processes is being developed. As an example, sandwich toast bread is investigated at laboratory scale. Using Design of Experiments (DoE), relevant influencing factors are systematically identified, while 3D scanning provides precise data on volume, shape, and structure. Based on these data, AI models are created to predict product quality and manage process variability. The objective is stable product quality despite variability by transferring artisanal knowledge into digital models.

Methodology, Partner Roles, and Outlook

NEURA GmbH leads the project as the domain expert and is responsible for needs analysis, methodological design, data quality, and the practical integration of results. Industry partners from plant engineering as well as a university complement the consortium with production infrastructure, raw materials, machine and process expertise, and the execution of experiments.

The feasibility study is funded under the ZIM program of the German Federal Ministry for Economic Affairs and Climate Action (BMWK). It establishes a robust data and methodology base for subsequent industrial research and development projects. The long-term goal is to gradually transfer the developed AI models to additional baked goods and processes and to make them applicable to other areas of food processing.