For people who use AI music generators: what do you actually need that no tool gets right yet?

I keep seeing new AI music generators launch, but I'm curious what real users actually feel is missing — not the marketing features, but the day-to-day frustrations.

If you've used tools like this (for videos, games, podcasts, or just for fun),I'd love to hear:

• What do you use AI-generated music for, and how often?
• What's the #1 thing that makes a generated track unusable for you?
• Do you need to edit the output (fix one section, swap an instrument), or is a
  one-shot result good enough?
• How much does licensing / royalty-free clarity matter for your use case?
• If you stopped using a tool, what made you quit?

I ask because I keep hearing two opposite things: some people just want "type a
prompt, get a song, done," while others say the output is only useful if they
can fine-tune it afterward. Curious which camp this community falls into.

Everyone is welcome to try out MuseGen.
www.musegen.ai
Genuinely trying to understand what people need.Would really appreciate your honest take.