Obviously you have multiple agents justify why they picked a certain response and then create another agent that picks the solution with the best justification.
touché
You could get the agents to output something structured and then use a deterministic test if you're worried about that.