Talking to the Market: Field Evidence on AI Chatbots, Digital Advice Utilization, and Retail Trading Outcomes

Background: Conversational AI chatbots are increasingly embedded in retail brokerage applications, yet empirical evidence on how chatbot-based advice influences investor behaviour in real-world trading contexts remains limited. Purpose: This study investigates whether perceived chatbot competence and trust predict digital advice utilisation, and whether such utilisation is associated with self-reported trading outcomes, with financial literacy examined as a moderator. Uniqueness: The study extends prior research by focusing on investor–chatbot interaction and actual utilisation of chatbot guidance among active retail investors, rather than evaluating algorithmic stock-picking performance or purely conceptual advisory models. Methodology: A cross-sectional survey was conducted with 300 active retail investors who had recently interacted with AI-enabled investment chatbots. Multiple regression models were used to test direct relationships and the moderation effect of financial literacy. Findings: Perceived competence (B≈0.32, p<0.001) and trust (B≈0.32, p<0.001) are independently associated with higher digital advice utilisation (R²=0.306). Digital advice utilisation is positively associated with trading outcomes (B≈0.32, p≈0.02). Financial literacy also shows a positive association with trading outcomes (B≈0.30, p≈0.03), while the utilisation×literacy interaction is not statistically significant. Implications: The findings suggest that brokerage chatbots may support more disciplined investing behaviour, though causal interpretation is limited by the survey design. Future work should integrate chatbot interaction logs with objective brokerage trading records to evaluate long-run behavioural and performance effects. Keywords: AI chatbots; Digital advice; Retail investors; Stock trading; Financial literacy; Trading outcomes