Smart Triage

Designing AI-powered triage for faster diagnosis and treatment

Quin | Amsterdam 2023 - 2025

Context

As part of Quin’s product team, I helped explore how AI could ease the growing pressure on healthcare. Specialists were overloaded, and patients often waited months for diagnosis and treatment. We wanted to understand where intelligent technology could make a real difference in clinical workflows.

Business challenge

In gastroenterology, IBS cases filled clinics with routine referrals and long waits for basic advice. Specialists spent hours on low-priority cases while complex ones competed for time. We set out to create an AI-supported triage tool that could make care faster and more focused, starting with IBS and later expanding to other fields like eye care.

My role

Building on early research and clinical insights, I helped translate findings into interactive prototypes that would support implementation and hospital buy-in. I worked closely with designers and developers to ensure everything was usable, feasible, and aligned with Quin’s design system.

Building on early research and clinical insights, I helped translate findings into interactive prototypes that would support implementation and hospital buy-in. I worked closely with UX , Service Designers and devs to ensure everything was usable, feasible, and aligned with Quin’s design system.

What we did

Together with specialists, we mapped referral journeys, explored explainable AI, and built and tested prototypes with developers. Each round of feedback brought us closer to a tool that felt intuitive, trustworthy, and ready for real-world use.

Researching current workflows and referral challenges.

Mapping user flows around specialist needs.

Exploring explainable AI for trust and clarity

Prototyping, testing, and refining in loops

High-fi prototyping for testing, implementation, and hospital pitches.

Continuously testing with real users

The result

The result

Quin’s Smart Triage Tool reduces waiting times, lowers costs, and helps patients reach the right care faster. It prioritizes urgent cases, automates routine referrals, and frees specialists to focus on complex diagnoses.

The user flow below shows how the tool structures each step of triage, from patient intake to personalized advice. It organizes data, highlights priorities, and supports confident, transparent decisions while keeping clinical control in the right hands.

Structured data access:

Cutting through data overload

Patient questionnaires powered the AI analysis but also created a flood of data. We needed to keep it useful, not overwhelming, by structuring information so specialists saw key insights first while full records stayed within reach.

Solution

We designed a structured overview that prioritized urgent cases and surfaced key information first, while still keeping full medical records one click away. The result was a triage view that let specialists focus on what mattered most, faster and with confidence.

Explainable AI analysis:

Designing for trust and clarity

Early on, we realized that trust would make or break adoption. Specialists wanted to know how the AI reached its conclusions, not just what those conclusions were. If they couldn’t trace the logic, they wouldn’t use it.

Solution

We made transparency a design principle. Each AI-generated diagnosis came with a clear breakdown of symptoms, risk factors, and linked sources. The logic unfolded step by step, giving specialists full visibility into every decision. With that, the tool moved from mysterious to meaningful. Something doctors could trust, verify, and build upon.

Keeping humans in control

Testing showed that specialists wouldn’t trust a tool that made decisions for them. They wanted support, not substitution.

Solution

We built an editing mode where they could adjust symptoms, risk factors, and diagnoses, keeping every AI suggestion flexible and editable. The result was a system that supported clinical expertise instead of replacing it.

Treatment

Personal care, powered by AI

Once the analysis was clear, the next challenge was personalization. Specialists wanted to offer tailored care, but drafting patient advice manually was time-consuming. The AI could help, but only if it felt like their voice.

Solution

We built a system where AI suggested treatment options and auto-generated draft recommendations, while specialists refined and approved the final message. This saved time without losing the human touch, turning data-driven efficiency into genuinely personal care.

Reflection

Designing explainable AI for healthcare demanded both precision and humility. Every decision carried ethical weight, because AI can be wrong, and in medicine that matters deeply. The complexity of diagnosis showed us that context is everything. Whether a symptom was considered alarming often depended on subtle combinations of factors that only a specialist could judge.

By linking AI insights to clear sources and reasoning, we aimed to build understanding, not blind trust. It reminded me that in healthcare, good design is not about replacing expertise but about supporting it with clarity, care, and respect for human judgment.


I’m proud of how far Quin has come in shaping safer, smarter healthcare technology, and I truly believe in the impact their products can make. You can see what they’re working on at quin.md.