i-medical · Radiology

An algorithm decides which radiologist reads which scan. The people running it need to see why.

i-medical needed a workflow orchestrator: a system that routes a pool of radiology exams to doctors based on subspecialty, availability, workload limits, deadlines, and revenue balance. I designed it from competitive analysis to developer handoff in a fast-track engagement.

Clienti-medical
My roleUX Designer
TimelineLate 2025 · 5–7 weeks
ScopeIA to key hi-fi
Engagement~85 hrs, fixed
3competitor platforms analyzed, none of which offer public access
5core workflows architected: exam queue, doctors, rules, delivery timing, analytics
6differentiation opportunities identified that no competitor documents

The situation

A radiology group runs on a queue: CT, MRI, ultrasound, mammography, each exam needing the right subspecialist before a contractual deadline. Assignment by hand doesn't scale. i-medical was building an orchestrator to do it by rule: doctor calendars, per-modality limits, insurance and public-health routing, urgency, and a scorecard formula balancing workload and revenue across the group.

The engineering logic existed on paper. What didn't exist was the interface where administrators configure those rules, trust the resulting assignments, and intervene when reality disagrees with the formula. Twenty-column-table territory. My kind of brief.

What I did

Competitive analysis without touching a single competitor

Enterprise radiology tools from Philips, Siemens, and Intelerad offer no trials and no sandboxes. I built the analysis from vendor demo videos, case study documentation, conference presentations, and academic publications, then separated table stakes from gaps. Six capabilities in i-medical's requirements, including per-doctor revenue tracking and insurance-based routing for European markets, appear in no competitor's documentation. That reframed the project internally: less catch-up, more lead.

Architecture before aesthetics

The scope deliberately spent its hours on information architecture and low-fidelity wireframes for the five core workflows, then applied high-fidelity polish only to the key screens stakeholders needed for alignment. The hardest design problem was assignment transparency: when a formula distributes exams, the interface has to show why each one landed where it did, or administrators will route around the system by hand.

FIG. 01 · Assignment rationale Redrawn abstraction · NDA-safe
MRI · NEURO · SLA 04:12 REMAINING ASSIGNED WHY DR. B SUBSPECIALTY MATCH CAPACITY 3/6 SLA URGENCY REVENUE BALANCE REASSIGN · WITH REASON MANUAL OVERRIDE STAYS VISIBLE IN THE AUDIT TRAIL, FEEDS BACK INTO THE RULES.
Every assignment carries its own rationale. Weighted factors are ranked, and overrides are first-class, not workarounds.
Worth admitting

This scope contained no user testing. That was a deliberate tradeoff for speed, made explicitly with the client, not an oversight. The design leans on established patterns from the competitive analysis and on stakeholder review instead. Validation with working radiologists is the obvious next step, and the design is annotated with the assumptions it should test first.

Where it stands

Delivered: the competitive analysis brief, information architecture, wireframes for all five workflows, high-fidelity mockups for the key screens, and an annotated handoff package. The engagement ran asynchronously against a fixed hour budget, with optional design QA during implementation.

The full case study is available on request

The detailed version with wireframes, the competitive matrix, and handoff artifacts stays off the public site while the product is unreleased. If we're talking about working together, ask.

Or email hello@studiogridline.com

Next case study

Emergency response, redesigned