Executive Secretary
XII Conferencia Internacional de Ingeniería Mecánica
COMEC 2025
VII Simposio de Logística y Gestión de la Calidad
Resumen
Abstract
Compliance with quality management standards and EU
regulations is presenting manufacturing companies with increasing challenges.
The aim is therefore to develop an AI-supported assistance system that combines
a Large Language Model (LLM) with a semantic knowledge graph to support
standard-compliant behaviour in an efficient, transparent and explainable
manner. Relevant standards, regulations and documents are automatically
captured using document upload, web scraping and optical character recognition.
The extracted content is structured in a semantic graph with entities and
relations.
A fine-tuned language model uses Retrieval Augmented Generation (RAG) to access
the graph and external sources and generate context-sensitive, comprehensible
answers. A user-friendly web front end enables visualisations and automated
notifications when standards change.
Fields of application include automated standards testing, support in
development and production, audit preparation and change management.
The project makes an innovative contribution to the digitalisation and automation of standard-compliant quality management by combining artificial intelligence and semantic knowledge modelling. It supports manufacturing companies in efficiently and reliably complying with standards, systematically improving quality and accelerating development processes for components and assemblies - especially in the context of complex product variants. It promotes the continuous improvement process and opens up new approaches for modular component development.
Sobre el ponente
Nadine Kaltschmidt

Discussion