Quick answer: Streamer audiences are bug-finding goldmines. A 90-minute structured bug bash with a streamer surfaces real-player workflows your QA can't.
Streamers play your game in front of an audience that includes domain experts you didn't hire. Tap that audience deliberately.
Recruit one streamer
Pick a streamer with audience overlap. Pay them; structure 90 minutes around your test areas.
Provide a tracker channel
Direct Discord channel for that stream's bugs. Engineers in the channel triage live; streamer's audience sees responsiveness.
Limit scope per stream
One feature area per stream. 'Boss fights and tutorial' is a stream; 'everything' is not.
Publish what you fixed
Post-stream patch notes credit the stream's findings. Goodwill compounds; the streamer's audience becomes invested.
“Player communities are testers. Streamers concentrate that into something you can engage.”
Don't make every stream into PR. The bash is for QA; the goodwill is the bonus, not the goal.