In manufacturing, accurate data is as crucial as the materials themselves. Brückner Maschinenbau, a market leader in film stretching with innovative solutions for various industries and materials, faced a common yet critical issue: an ever-growing database of over 500,000 parts where maintaining data consistency was becoming an increasingly laborious task.

For several reasons, including the ingestion of third-party data (from libraries, certificates, and specifications), duplicates, heterogeneous nomenclature, and missing change logs, working with data became increasingly error-prone, inefficient, and costly.

This challenge was a thorn in the side of highly skilled Brückner engineers, who often spent valuable time searching for the right materials, or missed essential change notices for components due to duplicates in the database. The initial goal was clear: improve data consistency and retrieval to streamline operations through automated data harmonisation and enable the engineers to focus on tasks with higher value.

Our team began by introducing Brückner to the specific benefits of customised Large Language Models (LLMs) for automating data quality. As a foundation, we gathered requirements through user interviews and data exploration. The solution was a bespoke application that enabled users to upload data and correct it with the help of LLMs to then reintegrate the refined results into the database. The key innovation was a user-friendly app that made powerful AI tools accessible for daily operations, creating tangible value for Brückner that reaches beyond the scope of this project.

The process wasn't just about deploying a solution but also involved continuous testing (technical and user acceptance) and development. We worked closely with Brückner’s data management team, ensuring the system was tailored to their specific needs and integrated smoothly into their workflow. The project was delivered in record time, with the Minimum Viable Product (MVP) achieving highly promising levels of accuracy.

Building on the successful collaboration, we are very proud to support Brückner in other projects along its transformation journey. For example, we have also helped enhance the predictability and control of packaging film quality in Brückner´s production process.

The high complexity of the production process, involving dozens of material types and hundreds of machine parameters, made it challenging to achieve desired packaging film quality properties reliably. We leveraged AI to reliably predict and influence film quality before production, reducing waste and improving overall efficiency.

Explainable AI provided data-driven insights into the production process, fostering a deeper understanding of the process as well as increasing transparency and trust in the AI models. Notably, AI models were capable of capturing key underlying physical mechanisms to make accurate predictions on product quality based on machine and material parameters. These insights will serve as the foundation for AI-based predictive assistants developed in the future.

Our partnership with Brückner led to the rapid design, development, testing, and deployment of successful solutions within weeks. These solutions not only address critical operational challenges but also strengthen Brückner´s position as a forward-thinking leader in its sector. For other companies facing similar data quality issues, the key takeaway is clear: with the right technology, approach, and partner, even the biggest challenges can be transformed into opportunities for growth and innovation.

For organisations that are, like Brückner, looking to deploy new technologies at a larger scale to solve similar problems, exploring AI-driven solutions is a step towards unlocking real benefits.

Do you want to discover more AI use cases that create real impact for your business? Visit our webpage ‘AI in the industrial value chain’.