Imperfect Automation Schema
A pattern of lower trust towards AI-based than human support agents. After a failure is experienced, trust is lowered and slowly builds back up after experiencing a number of correct recommendations from the support agent. This pattern does not differ between human and AI-based support agents, but the trust towards AI is continuously lower than towards human experts.