Well, Stored Procs can be an additional level of abstraction or a set of functions/methods if you look at the database like an object or service. This can be beneficial, since you can hide underlying implementation details and change it when need be without breaking the app (as long as you leave the interface, e.g. the stored proc parameters, alone). Also, if you can hide your table details behind a set of stored procs, no matter how someone gets access to your database, they'll only be able to interact with it using the methods you designed and wrote for it --> there's less risk of someone firing up Excel and hacking into your database.
On the other hand, Stored Proc require extensive T-SQL knowledge and you'll be spreading your business and app logic across two sets of code bases, which can be a downside. Anything you can do in a stored proc can also be done in ADO.NET with straight SQL statements, and it's not even slower anymore.
A further plus for Stored Procs is the fact that if something is wrong in your proc, you can fix it merely by deploying the proc with the fix - you don't need to touch your C# app. This can be a great plus in a "hosted" environment if you only get 3 releases per year - applying a hotfix by means of fixing an Stored Proc can be your live-saver then! :-)
All in all: there are lots of pros and cons for or against stored procs; I'd suggest, if you feel comfortable with writing T-SQL, why not give it a try and see how it works for you. If it feels to cumbersome or too inflexible or like too much work in the long run, you can always switch back to using straight SQL in your C# app.
Just my $0.02. Marc