I have been talking informally with Oliver Thomas and Lisa Robinson in Customer Support about ways to integrate our work with theirs and ways to learn from them what our customers are thinking and doing and having problems with.
I came up with the following ideas:

1. Using RAFT 3 rollout as a pilot of some sort. One thing that would be easy for us to do would be to provide a link in the footer that said "Create RAFT Help Ticket". It could easily pass along a ton of RAFT 3 data(browser, Operating System, User, current error if any, and any relevant RAFT data) to a specific RT queue. We could also just have some sort of feedback arrangement with CS (Lisa?) where once a month we got a report of ALL RAFT-related RT tickets.
2. Work with CS to mine the RT data for all RAFT/Cognos/Space/Data Warehouse tickets and again look at the report periodically to learn what is going wrong.
3. The metadata project could benefit from interviews with CS and again mining RT data for what kinds of questions users ask about metadata-related issues, where metadata in this case includes Business Actions (Create a PO, Annual Salary Review, Create a Budget, Approve a Requisition), the associated Roles, associated Reports, associated applications, associated DataWarehouse Packages and feed code and tables, etc.

Item 1: Creating Help Tickets that have technical information filled out for the user:
This development project would solve the following problem: Customer Support gets a bug report for a web application, but the user does not know or report on what browser, what operating system, and what page did the bug occur. forcing
The solution is relatively straight-forward: create a small JavaScript program that when called opens up a window with a small form prepopulated with browser version, operating system version, page the bug occurred on , and any text the application showed as an error message. The user then fills in just what they were doing at the time and clicks send. The form submits to cgiemail or another emailing service, which then sends it on to the customer support RT queue. The user gets an email back with the ticket number, as per usual.