For me the most interesting part is how AI can help start with building up complex Automation rules/workflows. Mostly it is taking some time to start new automation. If this would suggest a basic template (not only one but couple) from all of those available options that would be nice.
However on the other side when I think about all that things that do something for you when we look on the part of learning.. There is no better way of learning (Jira and other tools..) than doing this by yourself first from scratch..
Look how it is now.. For example so many certificate questions are based on knowing JQL .. if in a few years we would switch to "show me my open issues from project x" instead of "project = x and (assignee = y OR reporter = y)" then not sure if we are taking a big shortcut. How certificates questions would look like in the future if everything would be provided for us without much thinking?
Of course at the end we see the query generated by AI and we can verify it but still we are going more in a fancy direction that is for now a layer on top of a feature than a "add more features to JQL so that we can do more complex queries". If we look at converting human reading queries to JQL a mini version of that was already for years when using quick search in Jira - we just have now more key words that are recognized.
It looks nice and fancy but true AI should give us something before we even think about it and suggest changes to the things that we regularly do to optimize and improve that would make us continuously learn instead of getting only the result that would be accepted as a first "best option".
What do you think?