The Influence of AI Chatbots on Customer Satisfaction, Trust and Repurchase Intention

AI chatbots handle a growing share of customer service, and their effects on satisfaction, trust and repurchase intention depend on how well they resolve requests and on what happens when they cannot. This paper models the chatbot service encounter with first-contact resolution, availability of human escalation, anthropomorphic design and request complexity as inputs, satisfaction and trust as mediating responses, and repurchase intention as the outcome. A simulation of 1,000 encounters shows that resolution is the dominant driver, with standardised indirect effects on repurchase of 0.45 through satisfaction and 0.30 through trust; that an escalation path raises satisfaction mainly when resolution fails, so that repurchase intention among low-resolution encounters rises from 0.32 without escalation to 0.45 with it; and that anthropomorphic design raises satisfaction when the chatbot resolves the request but lowers it when the chatbot fails, since human-like presentation raises expectations the failure then violates. The paper interprets the results through expectation-disconfirmation and service-recovery theory and derives design principles: invest in resolution first, provide escalation that is visible and fast, and match anthropomorphism to capability. A field design is specified.