The filling a request process appears to be more complex than it is due to superfluous form fields. The information about authors, title and DOI has already been entered by the requester. For the service to be a viable alternative, it should be (and appear to be) extremely simple to fill a request.
Tracking down a paper is time consuming enough, even if you are sitting on the network with access to most journals. Don't make it look like people have to type in additional information manually.
At articleak (see my other comment), we only ask for the link to the paywall, and then if we could not get it from there (we make the http request via Tor for obvious reasons so it may fail some times) we ask you for the title of the paper, and that's all. So in the worst case you have to fill in two fields, but in most case, only one: the link to the paywall.
However, I would not be as hard as you seem to be with the pirate would made The Paper Bay. When you really need a paper and you already lost half an hour searching for it in the web ocean, it's okay to take 2 more minutes instead of 30 seconds to ask for it on such a service.
Tracking down a paper is time consuming enough, even if you are sitting on the network with access to most journals. Don't make it look like people have to type in additional information manually.