In the first category, you might want to return real-time information. For example, when called your executable might take a snap shot of the office coffee machine and return it as a GIF image. Alternatively, a program could log who has used it and when, or retrieve data from another source.
Executables that process the results of forms will usually just want to store the information returned; as well as a confirmation message back, perhaps you could send a fax just to prove that it has got through. Alternatively forms could provide a querying interface, so your program goes away and does a search on the given details and returns the results as an HTML message.
Note carefully that your program runs on the server PC, not on the user's computer. If you receive text back which is a command of some sort, then think of the security implications carefully; this should not be too hard as you as are writing the program and so this can easily do the filtering of dangerous requests.