PREDICTIVE LABOR MATCHING: UTILIZING AI TO SYNCHRONIZE SEASONAL LABOR SUPPLY WITH AGRIBUSINESS DEMAND IN NORTHWEST CAMEROON

Abstract

This study evaluated a worker-side AI-powered labor matching platform designed to address information asymmetry in post-conflict Northwest Cameroon's agricultural labor markets. A concurrent mixed-methods design was employed across 10 agribusinesses, tracking 350 workers (175 women, 175 men) over six months. Data were collected through surveys, interviews, focus groups, and usability testing, with quantitative analysis utilizing descriptive statistics, Chi-square tests, Mann-Whitney U tests, difference-in-differences, and panel regression, while thematic analysis was applied to qualitative data. Findings revealed that voice-enabled, offline-capable platforms significantly outperformed text-based alternatives: seventy-three percent preferred voice interfaces and eighty-seven percent required offline capability. Platform users experienced significant improvements: +1.0 additional days worked weekly, +2,400 CFA higher wages, and reduced wasted travel by twenty-three percentage points (p<0.001). Agribusinesses also benefited through reduced labor shortages (-1.9 days monthly), lower hiring costs (-16,400 CFA), and decreased harvest waste (-7 percentage points). The study recommends that technology developers prioritize voice-first, offline-capable design with multilingual support and informal network integration, while policymakers should invest in digital infrastructure and regulatory frameworks enabling worker-side platforms.