• There isn't a "everything run faster" set of changes.

    When you adjust indexes, you potentially adjust execution plans as well. This means some queries may run faster, some may not, choosing poor indexes. Assuming statistics are updated, you should be getting good plans, but with parameter sniffing and skewed distributions, there's not gross way of doing this overall for all your queries and indexes.

    The DTA should be used with a grain of salt. It doesn't look to efficiently pick all indexes, but it does a good job. However the DBA needs to weight the value of adding another index, making an include or filtered index instead, or ignoring the recommendation. Don't forget that every index you add impacts your insert/update/delete operations, and can slow them down.

    The way to do this is to make changes slowly and test. You don't care when x% the DTA says. It's a guess, not necessarily anything else..