Det tenkende bryggeriet – kontekstbevisst KI for smartere produksjon (NFR IPN tillegg)

Offentlig prosjektsammendrag

Modern breweries and fermentation-based industries collect vast sensor data, but much goes unused. Turning raw measurements into meaningful insights often requires manual sampling, lab work, and expert interpretation, consuming millions of work hours while valuable production tanks sit occupied longer than necessary. PLAATO, in collaboration with the research partner Simula, aims to change this by combining advanced sensors, cloud technology, and AI. Already providing plug-and-play fermentation sensors used in over 35 countries, this new step will make fermentation data easier to understand through natural language - essentially letting producers ask questions and receive clear, reliable answers. The project proposes three R&D challenges with associated methodology: * Data quality assessment of multimodal sensor data * Processing and annotation of time series data for interpretation by LLMs * Developing Model Context Protocol tools for LLM-enabled data insights