LLMs already do that but the advantage they have is being able to walk you through a plausible explanation for why they’re offering one position over another.
Using the output of a “decision” model without insight into the reasoning for a given decision seems very trusting.