A three-tier framework. Care-Enablement, Decision Support, and Autonomous Action — built to extend cleanly across every domain of healthcare AI, not just the slice that shows up on a claim form.
A single, portable test. The Clinical Inference Test — five questions that sort any healthcare AI tool into the category it actually belongs in, regardless of what the vendor calls it.
Real tools, tested against the framework. Including a cardiac imaging platform that explicitly says it doesn't make clinical judgments — and turns out, on inspection, to make one anyway.
Built for the people actually in the room
Entrepreneurs and builders — know which tier you're building before you scope compliance and go-to-market, not after. Healthcare leaders — a due-diligence test that checks a vendor's tool against what it actually does, not what it's marketed as. Regulators — a shared vocabulary for telling a routine administrative tool apart from an autonomous clinical actor, before you're blindsided by the next one. Policymakers — a starting model for reimbursement that doesn't treat every AI tool as the same undifferentiated risk.
About the Author
Rafael E. Salazar II, MHS, OTR/L is a healthcare strategist, advisor, and host of The Better Outcomes Show, where this paper's central case studies first took shape in conversation with physicians and healthcare technology leaders. He advises healthcare organizations and technology companies on positioning, go-to-market strategy, and clinical adoption — including AI tools that have gone through the exact validation questions this paper raises.
Read it before your next vendor pitch, board meeting, or regulatory filing.
Get a reliable framework for evaluating and implementing AI tools in healthcare.
Read the Paper Now!