Good question, Ganesh. For me it depends on how complex the problem is and how easy it is to reach a person.
I'm happy to start with an AI agent for quick, well-defined things (checking a status, a simple request, basic information), especially if it's clear from the start that it's AI and what it can do. I want a human straight away when the issue is complex, urgent, or involves money, security or an exception to the rules, or when the AI has failed to understand me once or twice.
To me the real line is the exit: if there's a fast, obvious way to reach a person, I'll try the AI first. If I get stuck in a loop with no way out, trust drops quickly, however good the AI sounds. Curious whether those building AVAs here measure how often customers ask for a human, and at which point in the conversation.
------------------------------
Robson
Coordenador de Infraestrutura
------------------------------