Impact of AI-powered document management on office productivity: An empirical study

Thesis Abstract

The study investigates how AI-powered document management systems influence office productivity in contemporary organizational settings, addressing the persistent gaps between digital tools adoption and measurable performance gains. Despite widespread deployment of AI-enabled features such as automated classification, semantic search, and automated workflow routing, evidence on their substantive impact on efficiency, accuracy, and employee satisfaction remains inconclusive, varying across industries and organizational maturity. This research aims to quantify these effects, identify mediating factors, and articulate conditions under which AI-driven document management yields significant productivity improvements. The primary aim is to determine the extent to which AI-powered document management practices affect office productivity, with specific objectives to (1) assess changes in task cycle times and error rates associated with document handling pre- and post-implementation; (2) evaluate perceived and measured productivity across knowledge workers and administrative staff; (3) examine the role of organizational factors (training, user experience, and process alignment) as mediators or moderators of productivity gains; (4) compare productivity outcomes across industry sectors and organizational sizes; and (5) provide a prescriptive model linking AI features to operational performance metrics. The study draws on the Technology-Organization-Environment (TOE) framework and the Job Demands-Resources (JD-R) model to explain adoption outcomes and worker well-being in relation to document-intensive tasks. A mixed-methods research design is employed, integrating quantitative and qualitative strands. The population comprises 42 medium-to-large enterprises across finance, legal services, and public administration within a metropolitan region that have implemented AI-powered document management within the last 12–24 months. A stratified random sample of 320 knowledge workers and administrative staff is selected, with a target response rate of 75%. Quantitative data are collected through system-generated analytics (e.g., average time to locate or process documents, document retrieval accuracy, workflow throughput) and a structured survey measuring perceived productivity, system usability, and training adequacy. The qualitative component includes 24 in-depth interviews with program sponsors, IT managers, and frontline users, complemented by 10 focus groups to capture contextual nuances and user-experience narratives. Data collection instruments are validated through pilot testing and expert review for content validity and reliability. Quantitative analysis employs descriptive statistics, paired-sample t-tests to compare pre- and post-implementation metrics, multiple regression analyses to identify determinants of productivity, and hierarchical linear modeling to account for nested data structures (individuals within teams within organizations). ANOVA is used to compare productivity outcomes across sectors and organization sizes. Mediation and moderation analyses test the indirect effects of training and process alignment on the relationship between AI features and productivity. Qualitative data undergo thematic analysis using a realist approach to extract patterns related to acceptance, perceived usefulness, and workflow integration, with triangulation performed against quantitative results to enhance validity. Expected findings indicate that AI-powered document management significantly reduces document handling cycle times and error rates, with moderate-to-high improvements in perceived productivity among knowledge workers, particularly where substantial training and process re-engineering accompany technology deployment. The study anticipates that user-friendly interfaces, robust semantic search, and automatic workflow routing will positively influence productivity, moderated by organizational readiness and task complexity. Sectoral differences are expected, with higher gains in finance and legal services due to stringent accuracy requirements and paper-to-digital transition intensity. The contribution to knowledge includes a robust empirical model linking AI document management features to measurable productivity outcomes, clarifying the roles of training, usability, and process alignment as critical determinants of benefits realization. Based on findings, the study recommends investing in comprehensive user training, aligning document management workflows with core business processes, and implementing continuous monitoring of system performance and user satisfaction to sustain productivity gains. It also suggests framework conditions for scaled adoption across sectors and outlines avenues for future research, including longitudinal studies to capture long-term effects and explorations of the environmental and organizational factors that sustain continuous improvement in AI-enabled documentation practices.