The FAIR Data Principles are widely applied to research data infrastructures, but assessing adherence to these principles remains challenging, especially for non-digital objects. The Research Data Alliance's FAIR Data Maturity Model (RDA-FDMM) proposes 41 indicators to measure FAIRness. We share experiences assessing KonsortSWD's PID service using RDA-FDMM and discuss automatic assessment using the F-UJI Tool, which employs RDA-FDMM and FAIRsFAIR Metrics in a machine-readable fashion.

The RDA-FDMM defines indicators, priorities, and evaluation methods for FAIR principles, organized into three classes (Essential, Important, and Useful) and five levels. We applied RDA-FDMM to KonsortSWD's PID service, aiming to assign PIDs to data elements below study level (such as survey variables). The PID service, an extension of the data registration agency da|ra , assessed some elements at the PID service level and others at the da|ra level. KonsortSWD's PID service assessment achieved high compliance with essential and important indicators.

The F-UJI Tool aims to provide automated FAIR assessment for research data from trustworthy repositories. It considers only indicators that can be assessed automatically, covering 16 of the 41 RDA-FDMM indicators. We used the F-UJI Tool to assess a GESIS dataset example and identified measures to improve FAIRness, increasing our research data score from 47% to 74%.

Automatic tools partially support FAIRness evaluation, as some aspects require human mediation and interpretation. However, tools like F-UJI are valuable for identifying weaknesses in metadata and metadata presentation that can be improved by automatic means.

The RDA-FDMM is a comprehensive standard for manual FAIR assessment recognized by the community and experts. Our experience highlights the importance of evaluating both machine-readable as well as non-machine-readable elements. Automated tools have limitations and technical challenges but offer valuable feedback for improvements. As the research ecosystem evolves, providing easily machine-readable metadata becomes increasingly important. We recommend adopting a "FAIR by design" approach early in product or service development to ensure FAIR principles are embedded in project outcomes and conducting regular FAIR assessments throughout the project lifetime to continuously evaluate and innovate research data infrastructures.