Artificial Intelligence Functions as a Complementary Capability for Monitoring and Evaluation System Effectiveness in African Public Management

Abstract

Artificial intelligence (AI) adoption in Monitoring and Evaluation (M&E) practice has outpaced empirical evidence on its institutional contribution to public management and development policy. This research examines whether deeper AI integration is associated with stronger M&E system effectiveness and whether more effective M&E systems correspond with greater Organizational Evidence Capability and Net Benefits. A cross-sectional analytical survey covered 75 M&E practitioners across 19 African countries. Empirical analysis preserved ordinal properties of survey measures and combined Kruskal-Wallis tests with rank-based associations. Cumulative-logit models provided adjusted estimates that accounted for M&E system maturity. AI-use intensity showed significant associations with System Quality (H = 42.26, ε 2 =.553, p <.001), Information Quality (H = 26.98, ε 2 =.338, p < .001), and Service Quality (H = 12.73, ε 2 =.137, p =.005). Adjusted ordinal models confirmed this pattern, with cumulative odds ratios of 2.75 (95% CI [2.00, 3.78]), 1.94 (95% CI [1.34, 2.82]), and 1.60 (95% CI [1.01, 2.55]), respectively. Stronger M&E System Effectiveness also showed substantial associations with Organizational Evidence Capability (OR = 10.82) and Net Benefits (OR = 7.15). Deeper AI use corresponded with greater reporting efficiency and improved data quality, while evidence use in organizational decisions also increased. No statistically significant association emerged for data-processing time. Findings support a conception of AI as a complementary organizational capability whose value depends on integration into established evidence processes rather than technological access alone. For public management and development policy, AI adoption should therefore remain anchored in M&E system performance and clearly defined evidence requirements for decision-making.