Implementation of an AI-Based Legal Assistant
Session
Computer Science and Communication Engineering
Description
This paper presents the development of a prototype AI-powered legal assistant aimed at improving the efficiency and quality of legal practice in Kosovo. The system integrates and processes public legal documents, including the Official Gazette, court rulings, and other official justice-related sources. Leveraging large language models (LLMs), RetrievalAugmented Generation (RAG), and complementary techniques, the assistant provides direct support for legal professionals by enabling faster and more accurate access to updated legal information. In addition, the prototype generates personalized legal texts—such as contracts or other legal forms—adapted to individual lawyers’ styles. By allowing practitioners to upload reference documents, the system can produce outputs aligned with their preferences, reducing time spent on drafting and enhancing productivity. The proposed solution represents a significant step toward the digitalization and automation of legal processes in Kosovo, contributing to a more efficient, transparent, and accessible justice system.
Keywords:
Legal assistant, artificial intelligence, large language models (LLM), RetrievalAugmented Generation (RAG), legal documents, personalization, automation, justice system, Kosovo
Proceedings Editor
Edmond Hajrizi
ISBN
978-9951-982-41-2
Location
UBT Kampus, Lipjan
Start Date
25-10-2025 9:00 AM
End Date
26-10-2025 6:00 PM
DOI
10.33107/ubt-ic.2025.69
Recommended Citation
Azemi, Agon; Qehaja, Besnik; and Hajrizi, Edmond, "Implementation of an AI-Based Legal Assistant" (2025). UBT International Conference. 1.
https://knowledgecenter.ubt-uni.net/conference/2025UBTIC/CS/1
Implementation of an AI-Based Legal Assistant
UBT Kampus, Lipjan
This paper presents the development of a prototype AI-powered legal assistant aimed at improving the efficiency and quality of legal practice in Kosovo. The system integrates and processes public legal documents, including the Official Gazette, court rulings, and other official justice-related sources. Leveraging large language models (LLMs), RetrievalAugmented Generation (RAG), and complementary techniques, the assistant provides direct support for legal professionals by enabling faster and more accurate access to updated legal information. In addition, the prototype generates personalized legal texts—such as contracts or other legal forms—adapted to individual lawyers’ styles. By allowing practitioners to upload reference documents, the system can produce outputs aligned with their preferences, reducing time spent on drafting and enhancing productivity. The proposed solution represents a significant step toward the digitalization and automation of legal processes in Kosovo, contributing to a more efficient, transparent, and accessible justice system.