I see what you mean, and I agree. But it's doubtful to me that we're close to a point where people would feel like a voice assistant is an entity worth having meaningful conversation with. It's still a dumb machine, but I'm sure that will change in the future - for some people.
Or perhaps I'm overestimating what people know about current 'AI' systems. The less you know about or understand a subject, the more mystical and sometimes interesting it appears.
Maybe the goal isn't to have people waste time talking to the voice assistant; rather, the goal is to use the voice assistant to discourage the chatters from burning human support agent time by making them navigate a machine system that filters for "real" issues.
Customer support is a cost center for most companies, so I can see where the strategy of a) minimize costs and b) provide just enough support that the FTC doesn't get involved while making the experience frustrating enough (including the cancellation experience) that most users accept crappy service.
Or perhaps I'm overestimating what people know about current 'AI' systems. The less you know about or understand a subject, the more mystical and sometimes interesting it appears.