Complementarities between AI- and Human Career Guidance in Kenyan Youth: Information, Recommendation, and Persuasion
Work in progress · collecting baseline data (Kenya)
Personalized career guidance is one of the few scalable tools that targets the information frictions behind occupational mismatch in low- and middle-income countries, but counselor time caps its reach. Can AI relax that constraint? I run a randomized controlled trial with about 4,000 young jobseekers across four urban and peri-urban counties in Kenya, comparing human-only, AI-only, and hybrid AI-plus-human guidance in a 2×2 factorial, and separately randomizing whether AI support stops at a recommendation or adds persuasion and action support. The design uses AI as a research instrument: it fixes the informativeness of advice and the strength of persuasion independently across arms, something a study of human counseling cannot do, and so separates whether guidance works by informing jobseekers or by persuading them to act. Primary outcomes are employment and earnings, match quality and persistence, and welfare, with short-run measures anchored to long-run welfare through a surrogate index estimated in external Kenyan panel data (KLPS).