Abstract
Operationalizing the EU AI Act requires clear technical documentation toensure AI systems are transparent, traceable, and accountable. Existingdocumentation templates for AI systems do not fully cover the entire AIlifecycle while meeting the technical documentation requirements of the AI Act. This paper addresses those shortcomings by introducing open-source templatesand examples for documenting data, models, and applications to providesufficient documentation for certifying compliance with the AI Act. Thesetemplates track the system status over the entire AI lifecycle, ensuringtraceability, reproducibility, and compliance with the AI Act. They alsopromote discoverability and collaboration, reduce risks, and align with bestpractices in AI documentation and governance. The templates are evaluated and refined based on user feedback to enableinsights into their usability and implementability. We then validate theapproach on real-world scenarios, providing examples that further guide theirimplementation: the data template is followed to document a skin tones datasetcreated to support fairness evaluations of downstream computer vision modelsand human-centric applications; the model template is followed to document aneural network for segmenting human silhouettes in photos. The applicationtemplate is tested on a system deployed for construction site safety usingreal-time video analytics and sensor data. Our results show that TechOps canserve as a practical tool to enable oversight for regulatory compliance andresponsible AI development.