The FDA defines SaMD as software intended for medical purposes without being part of a hardware medical device. AI diagnostic tools, clinical decision support systems, and patient monitoring applications may qualify depending on intended use and risk level. SaMD classification triggers a regulatory approval process before deployment.

AI agents in clinical settings that meet SaMD criteria must be FDA-cleared before deployment. Organizations that deploy uncertified SaMD face regulatory enforcement, patient safety liability, and reputational damage. The SaMD landscape is rapidly evolving as AI enters healthcare.

Healthcare AI agents submitted for BorealisMark certification receive additional scrutiny under the clinical AI audit framework. The audit includes SaMD classification assessment, bias testing across demographic groups, and hallucination risk evaluation in medical information contexts.