Differential ingests pathology, imaging, scripts and history. It generates ranked diagnoses with confidence, cites the evidence chain, and runs a self-critique pass to catch what a tired clinician would.
Differential surfaces the branch points. You see why it ranked cholecystitis over cholangitis, what investigation would break the tie, and which findings were weighed against each other.
Differential runs a dedicated red-flag pass on every case. Before you close the consult, it asks the question a senior clinician would ask at 3am — calmly, in writing, traceable.
Every claim below is measurable, repeatable, and traceable back to a case file. We publish our methodology.
“The critique pass is what sold us. It surfaces the second-order diagnosis I was going to miss — every time, without making me feel stupid.”
“I still sign every note. Differential just makes sure the note I sign is the right one. It's defensive medicine without the cost.”
No. Differential produces a ranked differential with confidence and evidence — a clinician always reviews, edits, and signs. It is decision support, not autonomous decision-making. Every output is a draft.
Low-confidence cases are flagged explicitly, with the specific investigations that would raise confidence. Differential will say “I don’t know, and here’s what would help me know” — we consider that a feature, not a bug.
No patient data is used for model training. Ever. All inference is run against our deployed reasoning engine with per-clinic data isolation. You can see exactly what was sent, when, and why, per case.
Socrates gathers history → Scribe captures the consult → both feed into Differential’s reasoning. You can also run Differential standalone against a patient file at any point. They compose; they don’t require each other.
Differential is clinical decision support, not a diagnostic device. We align with TGA guidance (AU), MHRA (UK), and are pursuing FDA 510(k) for US deployment. Security is SOC 2 Type II, ISO 27001, HIPAA-aligned.
Every case carries a reasoning trace, a confidence score, and a critique log. When a clinician overrides the ranking, we learn which class of case the model struggled with — and publish it quarterly. Transparency is the product.
45 minutes. Your referral, your labs, your imaging. You see exactly what the differential produces, how the critique pass works, and where the clinician stays in control.