Abstract
Artificial intelligence (AI)-driven personalisation increasingly underpinned social commerce interactions, yet evidence from African polytechnic student entrepreneurs remained scarce. This study examined how AI personalisation, operationalised as chatbot interaction quality (CIQ) on WhatsApp and recommender system quality (RSQ) on Instagram and TikTok, shaped loyalty intentions among students who actively traded via these platforms at Federal Polytechnic, Ede (FPE). A quantitative, cross sectional design was implemented. Structured questionnaires captured CIQ, RSQ, perceived value, trust, and loyalty intentions on 5 point Likert scales. Data from 412 valid respondents were analysed using covariance based structural equation modelling (CB SEM). The measurement model demonstrated strong reliability and validity (α and CR > .70; AVE > .50; HTMT < .85). The structural model explained R² = .441 of perceived value, R² = .418 of trust, and R² = .592 of loyalty intentions. RSQ exerted stronger effects on both mediators than CIQ, while trust emerged as the strongest direct antecedent of loyalty (β = .481, p < .001) relative to perceived value (β = .317, p < .001). Bootstrapped indirect effects confirmed dual mediation through perceived value and trust. The study concluded that, for FPE student entrepreneurs, AI personalisation functioned primarily as a trust building engine, with value augmentation as a secondary pathway to loyalty. The findings offered actionable guidance for resource constrained student businesses to prioritise transparent, reliable recommender and chatbot features to secure loyalty.
Keywords: AI Driven, Student Entrepreneurship, Chatbots, Recommender Systems, Consumer Loyalty.