Two design solutions
The first solution adapted EHR-style demonstration interfaces and integrated OnPoint AI summaries so the client could see how the AI might fit into familiar workflows. The second was OnPoint Iris AI: an experience gathering EHR information, organizing it by health issue, and surfacing relevant history, adherence gaps, missing tests, and comparisons with previous lab results.
Pre-visit: prepare with context
After an appointment is booked, Iris can assemble relevant patient history, previous medications, lab results, missed medications, and related health issues into a concise summary. It supports preparation; it is not presented as an autonomous diagnostic system.
At-visit: assist documentation
The intended workflow uses an Iris device to listen to the doctor–patient conversation and help populate relevant EHR fields such as medications and tests discussed. The clinician can approve, edit, or delete incorrect entries. Conversation retention was part of the intended workflow; implementation and compliance details require validation.
After-visit: review and correct
The clinician reviews the record, corrects incorrect entries, and manually adds missing information. Iris can compare current and previous lab results and summarize normal findings, items that may need urgent attention, and medication considerations for clinician review.
Discovery and collaboration
The work began with a briefing from the CTO and COO and product documents. I documented questions, sent a questionnaire, and joined a clarification call. After the first design presentation, a doctor, hospital staff member who manually enters patient information, and an experienced medical scribe contributed workflow knowledge. Iterations followed before the designs were frozen.
Design principle
The experience needed to make AI assistance useful and reviewable, keep clinicians in control, and reduce unnecessary effort without replacing clinical judgment.
Outcome
Interactive design demonstrations and a structured concept spanning pre-visit, at-visit, and after-visit workflows. This was a design engagement, not evidence of deployment in live clinical care or clinical validation.
My contribution
UX architecture, information architecture, user flows, UI design, prototyping, AI interaction patterns, design-system considerations, and client-facing presentations.