AI-Based Decision Support Systems and Administrative Efficiency in Saudi Public-Sector Organisations: The Mediating Role of Decision-Making Quality
DOI:
https://doi.org/10.51483/IJAIML.2946Keywords:
artificial intelligence; decision support systems; decision-making quality; administrative efficiency; task-technology fit; IS success model; public sector; Saudi ArabiaAbstract
Purpose: Governments are increasingly adopting decision support systems (DSS) based on artificial intelligence (AI), yet little empirical research has examined how these systems affect administrative efficiency. Drawing on an integration of DeLone and McLean's information systems (IS) success model and task-technology fit (TTF) theory, this study examines how the information quality, system quality and task-technology fit of AI-based DSS (AI-DSS) relate to administrative efficiency through decision-making quality.
Design/methodology/approach: Survey data from 455 employees and managers in 22 Saudi public-sector organisations that use AI-DSS in their work were analysed using covariance-based structural equation modelling with robust maximum likelihood estimation, 10,000-resample bootstrapping for mediation analysis, and model comparison procedures.
Findings: All three AI-DSS characteristics were positively associated with decision-making quality (R² = .545), with task-technology fit the strongest predictor (β = .407). Decision-making quality was the strongest predictor of administrative efficiency (β = .344; R² = .507) and partially mediated all three relationships, accounting for 37-39% of each total effect. Task-technology fit and information quality also had direct positive associations with efficiency, whereas the direct association between system quality and efficiency was weak and not robust to alternative specifications.
Originality/value: This paper shifts the focus of research on public-sector AI from adoption to post-adoption value realisation. It identifies decision-making quality as a mechanism through which AI-DSS characteristics relate to administrative efficiency, and it distinguishes a judgement pathway, through which system characteristics improve decisions, from an execution pathway, through which fit and information quality streamline task performance directly.
