As digitalization increases its pace, the legal field produces far more text than manual analysis can keep up with. Parsing large bodies of material such as Intellectual Property rulings by hand is both slow and labor-intensive, to the point where it is considered unfeasible. In this paper, the CLAIM tool, developed to tackle this challenge by blending conceptual modeling with a large language model, is presented and evaluated. Conceived as a Design Science artifact, the tool was motivated by stakeholder requirements gathered through participatory modeling workshops. The prototype runs in two consecutive stages: automated information extraction followed by semantic querying. Its capabilities are illustrated on a corpus of EU trademark-cancellation decisions brought on grounds of bad-faith filing. CLAIM’s evaluation was based on user-testing and interviews, resulting in a set of highly positive responses.