Sabrina Ramonov treats Claude like a new employee, starting with intern-level tasks and gradually expanding its responsibilities as trust builds. She restricts permissions at first, then grants broader access as both she and the AI learn to work together.
I think it's really like Claude's your intern at first and will work its way up to your CEO as you train it and it learns and continues on.
And it's a two-way thing too. Like as you learn how to partner and collaborate with it, you start to trust it more. So I'm like, well, in the beginning, would you give your new employee like access to everything and be like, go for it?
Like, probably not.
Like maybe a few things, maybe read-only access. Maybe they can't like update Stripe. And like billing information yet until they are like ramped up and you're confident, you know, that they can handle it. Same thing with AI tools. You can set permissions so that, let's say Claude can only read your data, it can't update it without asking for your permission. So like I, I really try to use that Menstrual Model of like, why, why are you trying to make it do everything on day one? Like it's gonna take some time and like you are in the process of like learning how to collaborate with it, learning what to trust it to do, and you're gonna find you don't want it to to do certain things. Like, there are a bunch of things I don't allow it to do, and that's how I like it. But I've learned that over time, so
