I don't exactly know what the nature of the content is, but the development of banner blindness depends on the extent to which the user finds the content routinely irrelevant.
When you first make the change, people will notice it. If it's something useful – say, a quick link to a document that people have to reference several times a day – then users are likely to appreciate the change. In such case you don't have to worry about banner blindness.
If the content is objectively useless – "always fill out your TPS reports!" –, then you're going to have a lot of issues. The only cure for banner blindness is increased relevance (better content) or novelty (new design, different content, motion, etc.), with novelty being less effective.
As for your to specific questions:
- Carousels are almost universally understood to be practically useless for anything other than impressing clueless clients with sparkles
- As stated before, the only sure cure for banner blindness is improve the quality of content such that it is interesting and relevant. Everything else is just rearranging deck chairs.
/edit
And in a very timely post Jakob Nielsen lets us know that people generally scroll past carousels, but lists some principles to use if you have to make them work: Designing Effective Carousels: Create a Fanciful Amusement, Not a House of Horrors