How does this result in less accountability, if for the AI agent to even have a chance to work properly it would basically need to save all of the calls, emails and other information in a persistent way, where it doesn't matter if it's hours or months since the last 'update'?
Just to be clear, this AI agent would be in the customer's control, nothing to do with whatever company it's dealing with.
> this AI agent would be in the customer's control
I misunderstood. In this case the AI can operate with the same limitations I can. It can save me some time when the interaction already had a good chance of success. But I can't see how an AI agent prevents the line from "accidentally" being disconnected, or being put in call waiting hell, or just having to talk to another AI in an endless loop (e.g. Google or Facebook's support).
True, it can't be saved from that at all. But it can still automatically keep a log of the whole process, instead of you having to do it manually, so it's easier to keep track what's going on for you as well. The AI agent can just automatically call back and wait for arbitrary amounts of time, as the assumption would be it has its own virtual phone number that can just keep chugging along without clogging up your actual phone.
If a company is essentially refusing to give customer service by hanging up or trapping callers in infinite loops, it's good to at least have a log of the whole process, without the customer having to risk their own sanity.
Just to be clear, this AI agent would be in the customer's control, nothing to do with whatever company it's dealing with.