Medlio
Medlio is an end-to-end redesign of a clinic and hospital management platform, rebuilt around the moment that matters most — the patient in the room. A fragmented, five-screen workflow reimagined as a single, focused consultation experience that turns a conversation into documentation in real time
About the Product
Medlio is a B2B practice management platform built for independent clinics and multi-department hospitals, where doctors, specialists, and admins share one unified workspace. Its defining feature is a live AI consultation mode that records, transcribes, and generates SOAP notes, billing, and referrals as the doctor speaks.
The product had the right idea and the wrong experience. Its core differentiator was buried four clicks deep, a single consultation was scattered across five disconnected screens, and the AI that justified the product's existence was a feature fewer than one in five doctors ever found. This was a complete rebuild — 37 screens across 8 user flows — designed to make the technology finally feel like it was working for the doctor instead of against them.
At its core, the Medlio redesign is about focus. It removes everything standing between the doctor and the patient, so documentation becomes a byproduct of the conversation rather than a task that starts after it.
Designing for Focus, Built for the Room
The most important decision in this project wasn't a screen — it was a reframe. The original brief was "modernize the interface." The real question turned out to be: how do we make documentation a byproduct of the conversation, instead of something that happens after the patient has left?
That single shift unlocked everything that followed. The consultation was rebuilt as one three-column screen — patient context on the left, the live session in the middle, the AI-generated SOAP note assembling on the right — so a doctor never has to think about the app while thinking about the patient. Every clinical action, from vitals to medications to referrals, lives as an inline modal inside that one screen. Five screens and fourteen interactions became one screen and four.
Visual Language
Medlio uses a calm, clinical visual language built for speed and certainty. In a medical context, ambiguity is a cost — so the interface makes system status impossible to miss. A live recording state with a running timer and waveform replaces a tiny corner icon. Section-by-section indicators show the doctor exactly what the AI has captured — green for ready, grey for pending — so trust in the output is built in real time, not requested at the end.
Restraint does the heavy lifting. Color is used to communicate rather than decorate: visit reasons and appointment statuses are color-coded for instant scanning, and allergy conflicts surface as high-visibility alerts at the point of prescription. Generous spacing and a consistent component system build muscle memory — Save always means Save, Send always means Send.Structured Storytelling
This structure ensures that users not only explore, but understand and connect with the content.
Structured Around the Session
Doctors don't think in software modules — they think in patients and sessions. The original navigation forced a clinical mind into a product structure: Billing, Referrals, Records. The redesigned information architecture begins instead with which patient, which session, and lets every downstream action flow naturally from there.
The work started with flow, not screens. Two weeks were spent mapping happy paths, edge cases, error states, and session-interruption scenarios before a single frame was drawn — and that's where roughly seventy percent of the design problems were actually solved. Distinct Live and Review modes give each phase its own focus: real-time transcription while the doctor listens, a centered SOAP note when it's time to verify.
A Foundation that Scales
Medlio isn't one layout — it's a system that holds up from a solo clinic to a multi-department hospital. The same information architecture serves an independent doctor's morning queue and a hospital admin's department view. Internal and external referral flows, clinic-level and hospital-level configuration, and a reusable component library mean the product can grow into new contexts without losing coherence.
Built for Real Use
A clinical product lives or dies on the unglamorous states. Empty queues, failed recordings, AI generation errors, dropped connections, first-time setup — every one was designed, not assumed. Billing auto-populates from the consultation's own orders and prescriptions, editable before sending, so doctors stop re-entering data the system already captured. Referrals happen inside the session as an inline panel, with doctor search, specialty filter, and a reason auto-suggested from the diagnosis.
Beyond the visuals, the redesign was built to ship. It went to engineering with a documented component library, a defined motion language, and a full handoff of spacing tokens and interaction notes — made to be implemented, not just admired.
Clarity that Scales with You
Every element of the redesign earns its place by serving the doctor's focus. Measured six weeks post-launch across 200+ active doctors: 38 minutes of documentation saved per doctor per day, a 34% lift in 30-day retention, a 3× jump in consultation-mode adoption now that the core feature is finally discoverable, a 62% drop in billing support tickets, and unguided task completion up from 1/5 to 4/5.
Get the question right, and the answers become obvious.
