This workflow is called Comfy Dancing as a tribute to the Dirty Dancing Movie that inspired me this workflow.
It takes a video of a person performing an action (it can be dancing, drinking, talking or whatevering) and an image of your influencer.
It will then make your influencer perform exactly the same motion as in the video using the new WAN 2.1 Fun model
The video motion can be extracted through 3 different methods that are explained in the Workflow, be free to experiment the 3 of them and see which one works best for you.
You can optionally isolate the character on the video if you want the action to be in another location (make sure to modify the prompt to explain which location you want)
You can as well activate the Faceswap group to perform a final faceswap to get sure your Influencer will be spot on :)
Enjoy!