Managed services became opaque for an understandable reason. The work is genuinely technical, the tooling is genuinely complicated, and it is quicker to say trust us than to explain a patch cycle. Over time that shorthand hardens into a relationship where the client cannot tell a good month from a bad one, and the provider has no particular reason to show them.
We think that arrangement is bad for both sides. A client who cannot see the work has no way to value it, so every renewal becomes a price argument. An engineer whose work is invisible ends up judged on how few tickets reached the owner rather than on how few problems existed.
So we report against published targets every month, including the months we miss them, we name every tool in the stack, and we hand back documentation on request without a conversation about it. None of that is technically difficult. It is simply a decision about who the information belongs to.