My question is after you write the initial user stories and then do months of "agile" back & forth with the business analysts/project managers how do you reverse engineer the actual requirements?
Our QA lead is constantly frustrated asking us "how is this supposed to work". I'm against keeping a giant word document with requirements b/c we would have to manually go and update it after EVERY ticket, yet that WOULD solve the problem.
Can anyone offer a more reasonable solution that would somehow leverage JIRA issues into a "living" document that would be a great reference for QA and even the devs with bad memories as they maintain the system? It can't just be r4j because all I see there are containers for tickets - I don't want to read multiple tickets - I want ALL the requirements captured accurately. Suggestions that work well for you?
There's no easy answer to this. I face this problem with many customers of mine.
First, you need to start building the documentation referring to the last release. Then, for every sprint, you need to prepare a new version of the documentation with the increments in the sprint. When the sprint is completed successfully and the code is merged, you have to merge the new documentation with the current one.
You can use Confluence for documentation, and on each update, you can reference the related Jira issues that caused the update.