CBR Intelligence · Research Paper
The Caribbean Cannot Afford to Get AI Wrong
But Can It Afford to Get It Right?
AI Stacking, Enterprise Competitiveness and Institutional Readiness
How can Caribbean enterprises capture AI’s productivity advantages without taking on costs and dependencies that exceed their capacity to generate value?
Joseph Cox examines the full cost of AI capabilities, the competitive risks of delay, and the skills, governance and institutional readiness needed to turn adoption into economic value.
Research abstract and principal findings
Abstract
This CBR Intelligence paper examines the economics of assembling and sustaining artificial intelligence capabilities in Caribbean enterprises. Using a 20-platform price register, illustrative technology architectures, a lead-to-proposal workflow model, enterprise cost scenarios and selected wage benchmarks for Jamaica, Trinidad and Tobago and Cayman, it evaluates conditional affordability and returns. It distinguishes access to AI tools from productive absorption and domestic value capture, and proposes enterprise governance tests and a sequenced regional policy agenda. The configurations are illustrative and do not represent observed technology purchases or measured adoption by Caribbean firms.
Principal research findings
- Full economic cost: subscriptions are only one component. Integration, prerequisite licences, skills, security, verification and vendor dependencies affect the cost of competitive capability.
- Returns depend on productive realisation: in the illustrative 120-case annual workflow, integration costs an additional US$1,380. At US$15 per economically realised hour, break-even requires 92 hours, or 38.3% of the modelled 240 hours released.
- Labour-value sensitivity: the same US$2,400 annual software expenditure requires about 41 realised hours at the selected Cayman professional-services salary benchmark, compared with 135–410 hours at selected Jamaican public-service minima and 119–318 hours at selected Trinidad and Tobago public-service salaries. These are conditional comparisons, not representative private-sector wage estimates.
- Architecture and institutional readiness: lean, modular, bundled and shared-service alternatives should be assessed against equivalent functionality, quality and safeguards. Accountable process owners, reliable data and clear decision rights determine whether faster work creates value.
- Regional policy: prioritise neutral stack-selection support, pooled evaluation capacity, incident sharing and proportionate governance over indiscriminate subscription subsidies.
These calculations are scenarios, not observed business outcomes. The full paper provides assumptions, technical tables, sensitivity analysis and references.
Research topics
Artificial intelligence in the Caribbean; AI stacking; enterprise competitiveness; AI affordability; total cost of ownership; productivity; institutional readiness; AI governance; SME digital transformation; domestic value capture.
Suggested citation
Cox, J. (2026). The Caribbean Cannot Afford to Get AI Wrong – But Can It Afford to Get It Right? CBR Intelligence. Caribbean Business Review. 21 October 2026.