Comment by stankalank

Combining AI and human expertise sounds like a robust approach. How do you balance the algorithmic recommendations with more 'human' relationship nuances?