Here's an example of the problem we face when trying to provide that information:
If Agent 1 and Agent 2 both use a single block of data in the repository (since it was deduped so that only one copy of it exists), how would we assign that space usage to each agent? Both Agent 1 and Agent 2 reference it and need it. If you delete Agent 1 from the repository and remove all it's recovery points, that block of data will still exist in the repository as used space (otherwise Agent 2's backups would be invalid). If we reported it as used space for both Agent 1 and Agent 2 that would double the amount of space actually used in the repository and would be incorrect (we're only keeping one copy of that block of data).
All we can do is report on the total amount of protected data (that's the raw data backed up for each agent prior to compression and dedupe) and report on the total used space in the repository. We can then calculate your space savings on the entire repository (we call it the compression ratio), but we can't provide you any calculations on space savings per agent.
The recommendation is:
Get a retention policy assigned to that agent and rollup enabled. Then during the next nightly job rollup will run and clean out much of that old data that is no longer necessary. If you track the used space in the repository prior to rollup and then after rollup you'll be able to see how much space was being used by those recovery points that were removed, but that's as close as you are going to come to being able to know how much actual disk space was used by a single agent in a repository.