Like I said above, we don’t use AI agents to grade other models. Instead, we run in-house evaluations tailored to each category of clinical AI, giving hospitals an apples-to-apples comparison between similar vendors.