The AI Matrix as Diagnostic: Access, Agency, and Adoption

This paper operationalizes the AI Matrix (Simpson, 2025), a two axis diagnostic of Access and Agency, using public summary data to support national AI strategy. Using public data from the IMD World Digital Competitiveness Report (WDCR) 2025 subfactor ranks (N = 69 economies), we construct Access (regulatory framework, capital, technological framework) and Agency (talent, training and education, scientific concentration) by inverting ranks so ‘higher means better’ when graphed. Adoption is proxied by IT integration. We place countries on the AI Matrix using median splits and test the Matrix’s central claim thatadoption rises with both axes but is capped by the weaker side. A falsification step searches for counterexamples (top quartile adoption with one axis below its median) and finds none. Correlations computed on inverted composites (Pearson) and as a rank based check (Spearman) show the strongest association between Adoption and the bottleneck proxy minimum (Access, Agency), not with an average or a difference. Robustness is probed by swapping the adoption proxy (cloud or ERP series), checking short run movement across adjacent years, and adding a light equity overlay (usage, trust, affordability) to guard against national average bias. We provide a minimal replication bundle and a Transition Playbook that translates Matrix position into near term policy actions. Limitations are noted (ordinal ranks, perceptual signal, within country heterogeneity). Even so, the AI Matrix offers a transparent, reproducible orientation tool for coordinated decision-makingand a bridge to further deeper causal analysis.