Getting the Human's Move

We would like human-move to behave like this:
     (display-game-state
       (human-move (make-game-state 5 8))) 
                                        
     Which pile will you remove from?       
     2                                      
                                        
     How many coins do you want to remove?  
     4                                      
                                        
                                        
     Pile 1: 5
     Pile 2: 4
The computer prompts the human twice for input:
(define human-move
  (lambda (game-state)
    (let ((p (prompt "Which pile will you remove from?")))
      (let ((n (prompt "How many coins do you want to remove?")))
        (remove-coins-from-pile game-state n p)))))