PracticeEHR

The check-in kiosk runs full screen on a touch device in the waiting room. Its core constraint shaped everything: most patients are older, many have low vision, and none of them came to the clinic to learn an interface.

Instead of a wizard with progress dots, check-in became a single vertical surface a patient can scroll like a paper form. Identity verification, license and insurance capture, demographics, and consent forms live in one continuous pass with large type, generous touch targets, and one clearly labeled action per moment.

Manual transcription of driver licenses and insurance cards was the biggest source of intake errors. The redesigned flow asks the patient to photograph both documents, runs computer vision over the images, and auto-fills the demographic and insurance fields for review.

The design problem was trust: patients need to see what the machine read and correct it without re-typing everything. Auto-filled fields are visually distinct, grouped by source document, and editable in place.

PracticeEHR had no CRM. I defined the information architecture from the ground up: how practices, providers, and patient relationships are modeled, what a relationship timeline looks like, and which reusable component patterns the rest of the platform could borrow.

The result is a system of list, detail, and timeline patterns that later product areas now reuse instead of reinventing.

The scheduling redesign centered on the Day Sheet, the screen a front desk lives in. A Wait List with Smart Match recommendations surfaces patients who fit newly opened slots, turning cancellations from lost revenue into filled appointments.

For the AI clinical scribe, which turns visit conversations into structured notes, the design is built around review speed: providers skim, correct, and sign rather than write. And the platform's legacy Power BI reports were rebuilt as modern dashboards a practice manager can read at a glance.