I'm moving to an ai search rag for my knowledge base. You just need to visualize the backlog by ticket category, type, activity and you can query all that in JIRA, wouldn't think a LLM is needed here unless you want to use it for classification, to add some labels, do typical github actions bot things.

The jira search syntax can be a bit much, as long as your tickets are sanitary then the expectation is you can make a few bar charts, and maybe evaluate ticket quality with a rating using a LLM. A lot of tickets were filed as todos, with no context detail, so you can order by body length, etc.

I have seen some backlogs by now, other than a maintenance culture question, what you want to do and what you can do without other people involved are usually two things. If fixing a bug raises complaints as to what people commit to in sprint, then the backlog is just an example of people that learned better than wasting time on backlogs.