HUCAI-S4 project is progressing in the implementation of its technological proposal: an artificial intelligence platform integrating human knowledge into its prediction capabilities to optimize industrial production processes. This innovative solution aims to improve efficiency, reduce errors, and increase the competitiveness of industrial companies through a platform that can be adapted to different production environments.

The consortium, comprising ZEO Technology, NAITEC, Bacaicoa, and Borg, has already identified the two real-world use cases on which the platform will be developed. In addition, multimodal data has been collected and analyzed, and the development of artificial intelligence models has begun.

In the case of Bacaicoa, the platform will be implemented in an automated process, where the aim is to predict the response to a quality test. To this end, the first step has been to collect and label data from both production and test results.

In the case of Borg’s application, where manual processes will be used, the aim is to optimise production planning. NAITEC and Borg have collaborated in the development of an intelligent sensing system to predict the condition of a component before its reuse. This system has been developed by integrating cybersecurity criteria in accordance with CRA regulations from the design stage onwards.

In addition, ZEO has focused on preparing the most efficient data capture, processing and architecture environments for AI exploitation, including data from both machines and people (plant operators, quality personnel, etc.), even in natural language, with the aim of digitising the knowledge of what is happening as a source of high-value information for predicting inefficiencies.

In the upcoming months, the consortium will focus on developing and validating predictive models and integrating human feedback to improve these models through continuous learning.

 

HUCAI-S4 in the context of current industrial AI solutions

HUCAI-S4 differs from other solutions in key aspects such as its ‘Human-in-the-Loop’ approach and multimodal artificial intelligence.

The first difference is the continuous interaction between humans and artificial intelligence, in two ways: the recommendation system based on AI predictions and human feedback on the quality of those recommendations.

This model allows operators to feed data into the system to correct and adjust AI predictions. This means that data sets are not limited to information from machine controls and sensors. In industrial production environments, large amounts of data are recorded manually or with a delay by humans and can be key to understanding process behaviour. Furthermore, this data can be in different formats (numerical, images or natural language), which often makes it difficult to use.

Secondly, HUCAI-S4 incorporates continuous learning into the prediction model. Once the AI has made its estimate, false positives or negatives may occur. Allowing people to provide feedback to the model, “teaching” it when it is wrong, is a valuable source of information for improving the accuracy of the system over time.

Another differentiating factor is the quality and relevance of the data. While many AI solutions rely solely on data volume, HUCAI-S4 places special emphasis on data quality, ensuring that the data used to train the models is representative and directly related to the problems to be solved. This approach also takes into account the knowledge of the people who are in direct contact with production, facilitating the digitisation of this knowledge and its integration into the platform.

 

About HUCAI-S4 consortium

HUCAI-S4 is coordinated by ZEO Technology, a technology company specialising in MES and OEE digital platforms, which is leading the design and development of the platform. NAITEC, as a technology centre, contributes its expertise in the development of advanced predictive models. Borg Automotive, the European leader in the remanufacturing of automotive components, and Bacaicoa Industrias Plásticas, which specialises in the production of three-layer film, provide real industrial cases to design and validate the platform in real production environments.

The HUCAI-S4 project was finalist in the Factories of the Future Awards 2025, in the Best R&D Project category, and has also been nominated in the Advanced Manufacturing Awards 2025, in the Most Disruptive Technology in the Production Process category.

The project has been funded by the Government of Navarra as part of the 2024 call for strategic R&D projects.