- The Story
MediScreen began life as a remote–candidate-screening platform: asynchronous video interviews, AI-generated transcriptions, and a slick dashboard for recruiters. Then March 2020 arrived. Hospitals were overwhelmed, clinics had no safe way to triage walk-ins, and patients were terrified of crowded waiting rooms. MediScreen’s founders saw an unexpected path forward: repurpose their one-way-video engine as a contact-free COVID-pre-screening tool that any GP practice could deploy overnight.
They turned to Vietne—the same studio that had designed their recruitment UX—to refit the product for healthcare at break-neck speed. The mandate was simple and audacious: keep the two-tap ease recruiters loved, but add medical-grade trust, symptom recognition, and AI analytics that could clear—or escalate—a patient before they ever left home.
From Hiring Portal to Pandemic Triage
When the virus hit, MediScreen’s founders didn’t scrap their recruitment engine—they re-fitted its core workflow for healthcare. The skeleton stayed the same (record ➜ upload ➜ review), but every supporting detail pivoted from HR to clinical safety:
Replaced body-language scoring: the model now checked for flushed skin tones, laboured speech, and coughing bouts.
Swapped behavioural prompts for symptom trees—fever, travel, loss of smell—updated daily from WHO guidelines.
Supplanted recruiter star-ratings: clips arrived in GP inboxes tagged green (routine), amber (monitor), or red (tele-consult now).
Design Challenges
Patients of all ages had to start a session in seconds, yet the interface needed the gravitas of clinical software.
In a pandemic surge, a GP can’t watch 200 full videos. We had to surface the urgent 10 % instantly and let AI summarise the rest—turning a raw clip queue into an actionable triage list.
Healthcare buyers care about minutes saved and admissions averted, not codec acronyms. The challenge was to lead with impact metrics while still giving IT teams easy access to encryption details, model lineage and uptime SLAs.
The Vietne Approach
A magic-link SMS opens a web view; if the camera isn’t centred on the face, a friendly overlay guides positioning. Average start-up dropped from 75 seconds to 22.
The moment a clip uploads, on-device AI scores symptom severity. Doctors land on red cases first, with cough keywords and oxygen-effort markers highlighted in the sidebar.
We rewrote every headline to speak in clinic hours saved and admissions averted. Deep specs—encryption, accuracy, model lineage—live behind accordion drawers for power readers.
Impact (First 8 Weeks)
Human Snapshot
Lina, forty-seven, woke one Thursday with a dry cough and slight fever. Instead of driving to her crowded clinic, she tapped the MediScreen link her GP’s office texted overnight. Thirty seconds of guided recording and four yes-or-no questions later, the system flagged subtle breathlessness and routed her clip to the top of the doctor queue. Within the hour she had a video consult, a PCR test booking, and clear isolation instructions—no waiting room, no panic, no hospital visit.
