Relationstool structures civil court proceedings in minutes
The Relationstool for case file analysis sets a new standard for efficient civil case management-and brings legal precision to practice faster with AI. Piloted with the Ministries of Justice in Schleswig-Holstein and Bavaria, the tool demonstrates that speed and legal accuracy go hand in hand. The “Law as Code” approach combines the rule-mapping method with modern language models-transparent, scalable-and implementable in days rather than months.
Challenge
Modern language models (LLMs) excel at analyzing texts and extracting content. However, they often fail when confronted with the strict, rule-based logic of legal subsumption. Civil courts are generally confronted with enormous volumes of text. The Relation Tool is designed to be cluster-agnostic: its effectiveness is demonstrated regardless of whether the cases involve complex individual cases or mass proceedings with identical factual circumstances. It structures every statement of facts precisely according to legal relational logic, without relying on predefined case groups.
A particularly critical issue for the judiciary here is how to manage these volumes of text without relinquishing legal control. Traditional automation approaches fail either due to the complexity of unstructured texts or the lack of transparent, traceable subsumption logic.
Solution
The answer lies in combining technologies. The Relation tool uses the rule-mapping method to deterministically define the legal logic, while AI handles the tedious work. The tool fully automatically analyzes all pleadings in a proceeding (complaint, answer, reply, rejoinder, etc.), identifies the relevant basis for the claim, and structures the parties’ statements of fact accordingly.
Impact
The result: a structured case file overview, a tabular mapping of the facts to the applicable legal provisions, a summary of undisputed and disputed points, and a fully drafted statement of facts. For judges, this means a clearly structured foundation for efficient legal work.
Commissioned by the Ministries of Justice in Schleswig-Holstein and Bavaria, the judicial assistance system was successfully developed and tested for civil proceedings. The tool acts as an assistant, not as a judge-decision-making authority remains with humans.
The principle is transferable: what works in civil litigation also proves effective in administrative courts. The combination of rule-mapping logic and AI-powered text processing enables scalable, transparent, and constitutionally sound digitization-implementable in days, not months or years.
