For a considerable amount of time, crossword puzzles have been widely played as language games that foster the growth of intelligence and test players' vocabulary, problem-solving skills, and knowledge. These puzzles were meant to be created through a process of finding and accommodating words into a grid manually. Of course, it should be ensured that these words fit together and do not conflict. Nowadays, there is new potential to automate and enhance the crossword generating process, as artificial intelligence continues to expand and evolve. Quite some studies have been performed to find the most suitable algorithm, but none of them in Albanian language. This study explores the potential and methods of artificial intelligence in solving these issues, focusing on the use of the Backtracking algorithm with an Albanian dataset. Creating and integrating artificial intelligence algorithms is the primary emphasis of the research. The reason the backtracking algorithm was used is because, when combined with constraint satisfaction issues, it explores every possible way to meet the requirements and produce an accurate and complete crossword. Moreover, this research dives deeper into the analysis of the many data sources of Albanian words, from which the dataset will be created.