The prescription, drafted before the visit ends.
A decision-support module that reads symptoms, surfaces likely conditions, and recommends medications in each doctor's own prescribing pattern.
Decision support
What they were up against.
A routine diagnosis takes a doctor 10 to 15 minutes, and a meaningful slice of that is the write-up: matching symptoms to likely conditions and composing a prescription. Done dozens of times a day, that adds up to real time lost at the bedside.
The goal was not to replace clinical judgment but to hand the doctor a strong first draft, informed by what this specific doctor tends to prescribe.
How the approach works.
The module takes the symptoms a patient presents, surfaces the conditions those symptoms most likely point to, and then recommends medications drawn from that doctor's own prescribing history. The personalization is the key move: it learns each clinician's pattern rather than imposing a generic formulary.
And it learns fast. After watching just 10 prescriptions, it begins tailoring its recommendations to how that doctor actually practices.
What it changes.
The module shortens the slowest part of the visit, the write-up, by drafting condition and medication suggestions the doctor can accept or adjust. Crucially, it gets personal fast: just 10 prescriptions is enough to start matching a doctor's own patterns.
Less time composing, more time with the patient in front of them.