Maybe I’m wrong, but you still can’t be 100% sure in any case. Even if you have your own support team, you still can’t control every specialist 24/7, and you’re not always aware if they told a customer some nonsense. You can build a system where the team almost never falls out of context. You need a single constantly updated knowledge base that both people and AI rely on, especially if this is
outsourced customer support. Plus you need regular quality control and review of conversations, otherwise even a good team will start answering superficially. And most importantly, you need to make sure there is a closed feedback loop, so new customer issues immediately turn into updates of processes and answers. Something like that.