Magic+ is actually not just one person, but it's a software-driven service with a highly trained team of professional assistants that are held to a very high bar. They are managed, supervised, and trained by us so that you don't have to worry about it.
And, on top of that, every task and request is managed from beginning to end using custom software that is designed specifically to monitor the execution strategy and enhance both the quality and the reliability of the output. It's difficult for me to say more right now about how this works without disclosing IP that we cannot disclose at this time, but I hope to be able to say more about how this works in the future.
A good way to think about Magic+ is that it's a natural-language command-line interface for complex tasks that require some human interaction. A component of how we handle your email is more like a ~/.procmailrc file that can pipe things to a highly-supervised human assistant than it is like having a random person somewhere in the US logged in to your Gmail. You tell us how you need things done, using natural language, and it gets done, assuming it's possible.
Again, I recognize that this explanation is not as detailed as it could be, but I hope that it helps. I will think about how we can explain this better before we are able to disclose the inner-workings of our software.
I do understand how the software layer plus having multiple assistants working on a single person makes it "safer" than the SPOF of a traditional assistant (and I'm sure you've already thought of ways to have the software handle all PII while the assistants just support with the bits that require manual labor)
However, I feel like the killer feature of this kind of product should be Trust. That's worth paying for.
That plus the minimalist website copy and secrecy around operations makes it seem a bit niche for the startup crowd. And that sucks, because I want the AI assistant winner to be a newcomer, and not the usual suspects :)