The links above probably explain better, they are screencasts that have the same or similar idea to what I am thinking of. Sorry if you have already seen these before.
For (a very simplified) Example:
Suppose you are writing an email using your regular desktop email client. Inside that email client there is probably a feature to configure "signatures" that go at the bottom of your email message. Moreover, there are probably multiple signatures that you can choose from, depending on the context.
The problem is, what if you want to send an email message using your web-based email client? Do you have to go in and re-configure the signatures again? What if your web-based mail doesn't support this feature?
One approach to this situation is to open up your desktop email client, get a signature out, use the windows "clipboard" and copy and paste from one application into the other.
Why not have a *centralized* place to put your email signatures? FARR can do this kind of thing easily. In addition, it would be great if you could have a single set of FARR commands to do all kinds of "message related" stuff without having to use fifteen different user interfaces to do essentially the same thing:
(the nice thing about this, these commands are kind of verbose, but FARR can already "autocomplete" and "auto-suggest" as you type, this means you don't have to memorize all the little arguments and stuff, and you don't have to literally type every keystroke to get this output)
This could not only "cut and paste" your content, but actually be smart enough to know which application you were sending the command to, and therefore use features that were sensitive to the context. For example, know which anonymous screen name to use, depending on which forum you are writing your message in.
This is just a simple snapshot. There are thousands of ideas that can be built of of this basic premise.