I'd like tags to be suggested based on machine learning, so I don't have to come up with rules, I work on many small projects, so creating rules manually is too time-consuming.
I'd like some machine learning to look at past tagging and suggest the correct tag (tags can be those found via the Jira integration, rather than AI making up text)
The machine learning should take into account adjacent tags, and how I've tagged things in the past. there should be a huge wealth of data.
It could also use the screenshots, often a client logo appears on those.