A conversation that collects the parameters, a fixed schema that holds them, and a rule engine that reasons over them. The agent never invents the diagnosis - it runs the rules and shows its work.
Works the case through with you. Opens on the presenting sign, branches its questions, and clarifies ambiguous wording until the parameter set is complete enough to reason over.
Runs the parameters across a defined set of nervous-system sites, scoring each at three confidence levels and recording which parameters drove the score. Differentials follow from the localisation; the work-up list is what's left to check.
History, mentation, gait & posture, pain, paroxysmal events - a fixed clinical schema.
Across a defined set of nervous-system sites, each at one of three confidence levels.
For every localisation, which parameters supported it - so you can check the reasoning.
Derived from the localised site and the same parameters, not from a free-form guess.
The in-clinic checks and tests still needed to narrow the differential.
Raised plainly when the picture is acute and worsening.
The agent doesn't decide the case. It localises, lists, and reasons - out loud - and leaves the judgement to you.See what it does → Read the pitch