Artificial Intelligence Applications for Document Management: Experimentations, Potentialities, and Challenges

Since 2022, with the launch of ChatGPT, Artificial Intelligence (AI) has gained prominence, popularity, and applicability across various sectors. Document Management (DM) is among the domains that can benefit from this computational solution. In this context, this study seeks to determine how AI can be employed in document management. The overall objective is to analyze the applicability of Artificial Intelligence to Document Management. More specifically, it aims to identify scientific literature addressing the intersection between AI and DM; to identify relevant works, authors, and groups for studies involving AI and DM; to identify AI methods, techniques, and applications for document management; and to outline theoretical frameworks, perspectives, and research trends. To this end, a multi-method approach is adopted for reviewing scientific literature. The procedures consist of four phases. In the first phase, a search was conducted in the Scopus database for the period from 2013 to 2023 using the query: ("document management" OR "records management" OR "electronic document management") AND "artificial intelligence". In the second phase, the data collected from Scopus were analyzed with the VosViewer tool to perform a quantitative and bibliometric analysis. In the third phase, the titles, abstracts, and keywords of the articles retrieved from Scopus and Consensus were examined to select a valid bibliography for the topic under study. After applying inclusion and exclusion criteria, the selected articles were read in full during the fourth phase of the project, which involved Content Analysis (CA). Additionally, the Consensus tool, an AI-based academic research platform, was employed to supplement the review based on the project's research question. As a result, the study presents a bibliometric analysis of 363 works identified in Scopus, including assessments of co-occurrence, citation, and co-citation patterns, as well as the qualitative analysis of 26 articles. A publication peak was observed in 2023, reflecting a rising trend since 2017, particularly in the United States, with healthcare being a notable area of focus. The categorization of articles also revealed approaches, technologies, potentialities, and challenges associated with employing AI in the field of Document Management, thus providing a roadmap for those interested in further investigations on this topic.