Manz, a legal publishing house in Austria, built an AI-based recommendation system to speed research workflows for lawyers and other legal professionals by leveraging semantic document similarity using deepset Cloud. Their online database, RDB Rechtsdatenbank, now offers improved search functionality with recommendations for semantically related documents, reducing the time invested in searching through hundreds of documents. The solution was built on top of a German BERT language model fine-tuned to the legal domain, and its integration into the database was straightforward once basic software development experience was possessed by the team. The new feature has improved user experience with more pertinent results compared to an earlier competitor's implementation.