Mm.

And because there's quite often a kind of misconception, of course, we have our act- our Active Directory, we have to take care, uh, having some backups for that one, yeah? But most of the things are not running on Active Directory.

We have our Entra, we have our Okta, we have our Ping Identity, um, cloud service somewhere, and we have a c- cloud service for that because we don't want to take care of it, yeah? Um, there's a provider for that, right? And, um, the misconception is very often, okay, what are the areas this provider's taking care of? What's the tasks, responsibilities the provider's taking care of? But what is the area I'm responsible for? And when it comes to configurations, but especially...

So, for instance, configurations like, hey, these are applications, uh, these are service accounts, but also content like, like users, like passwords or other types of credentials, yeah? Then quite often, um, quite often this is something where the provider's not taking care of, yeah? So it's, he's providing the robust infrastructure, but you are in charge of the configurations of the content, yeah? And also being able to recover it, yeah? So, um, to that state you want it to have to be.

And this, this is an area which is quite, um, quite, quite important.

So having a very, very crystal clear understanding of the responsibilities of the topic, and not just by, um, discussing it somehow, but also enforcing the participants to test it, to verify it, yeah? Um, which is a second very important element.

But, um, there are other areas, other areas as well.

But how are AI agents different than service accounts in particular or, or just non-human identities in general? What's-- What are your, uh, thoughts on it and recommendations on it?