Redesigned neuroimaging workflows to empower faster, AI-assisted clinical decisions
Redesigned neuroimaging workflows to empower faster, AI-assisted clinical decisions
Redesigned neuroimaging workflows to empower faster, AI-assisted clinical decisions
To transform BrainSightAI's fMRI diagnostics into a transparent, clinician-first experience that accelerates stroke and tumor detection while ensuring radiologist oversight and trust.
To transform BrainSightAI's fMRI diagnostics into a transparent, clinician-first experience that accelerates stroke and tumor detection while ensuring radiologist oversight and trust.
Designing an AI assistant that supports, not distracts.
Designing an AI assistant that supports, not distracts.
BrainSightAI’s AI tool helps radiologists interpret brain maps faster.
BrainSightAI’s AI tool helps radiologists interpret brain maps faster.
Clinician feedback revealed difficulty in interpreting AI-generated data due to dense visuals and scattered insights. Product teams observed confusion around key risk indicators, while founders emphasized the need for a clearer decision-making flow and improved trust in AI recommendations. These insights drove a redesign focused on visual clarity, guided workflows, and contextual interpretation of AI outputs.
Clinician feedback revealed difficulty in interpreting AI-generated data due to dense visuals and scattered insights. Product teams observed confusion around key risk indicators, while founders emphasized the need for a clearer decision-making flow and improved trust in AI recommendations. These insights drove a redesign focused on visual clarity, guided workflows, and contextual interpretation of AI outputs.
To redesign BrainSightAI’s neuroimaging platform into a clinically intuitive, AI-assisted diagnostic tool that enhances visual clarity, builds trust in AI outputs, and fits seamlessly into radiologists’ workflows—enabling faster, more confident stroke and tumor diagnoses.
To redesign BrainSightAI’s neuroimaging platform into a clinically intuitive, AI-assisted diagnostic tool that enhances visual clarity, builds trust in AI outputs, and fits seamlessly into radiologists’ workflows—enabling faster, more confident stroke and tumor diagnoses.
Exploring tools like Gleamer, Radmat, Medic-AI, and Brainsight's user journey revealed key pain points: cluttered visuals, scattered workflows, and low trust in AI outputs. Radiologists needed easier comparison between raw and AI-interpreted scans, clearer patient data access, and more control over AI suggestions. These gaps shaped a redesign focused on clarity, guidance, and user autonomy.
Exploring tools like Gleamer, Radmat, Medic-AI, and Brainsight's user journey revealed key pain points: cluttered visuals, scattered workflows, and low trust in AI outputs. Radiologists needed easier comparison between raw and AI-interpreted scans, clearer patient data access, and more control over AI suggestions. These gaps shaped a redesign focused on clarity, guidance, and user autonomy.
Focused on understanding the workflow: from patient ID collection → document upload → diagnosis review.
Focused on understanding the workflow: from patient ID collection → document upload → diagnosis review.
Timeline:
Timeline:
24 March, 2025 - 26 March, 2025
Nov 2023 - June 2024
Tools:
Tools:
Figma
Figma
Loom
Loom
Disciplines:
Disciplines:
Disciplines:
Product Strategy
Product Strategy
Product Strategy
UX Research & Design
UX Research & Design
UX Research & Design
UI Design
UI Design
UI Design
FOCUS
FOCUS
FOCUS
SCENARIO
SCENARIO
SCENARIO
CHALLENGES
GOAL
GOAL
USER REASERCH
USER REASERCH
SOLUTION
SCAN ANALYSIS
CLINICAL COPILOT
PROGRESS TRACKER
LEARNINGS
INSIGHTS
The redesign ensures radiologists remain in control, while AI acts as a reliable co-pilot — quietly assisting with smart insights, helpful suggestions, and accessible context — without disrupting the clinical workflow.
The redesign ensures radiologists remain in control, while AI acts as a reliable co-pilot — quietly assisting with smart insights, helpful suggestions, and accessible context — without disrupting the clinical workflow.
AI interprets fMRI scans in real time, highlighting high-, moderate-, or low-risk brain regions through color-coded overlays and confidence scores. Clinicians can customize regions of interest and export detailed diagnostic reports with ease.
AI interprets fMRI scans in real time, highlighting high-, moderate-, or low-risk brain regions through color-coded overlays and confidence scores. Clinicians can customize regions of interest and export detailed diagnostic reports with ease.
The built-in assistant translates AI findings into clear clinical language, flags signs of concern like stroke or tumor patterns, and suggests next steps—empowering doctors to act with clarity and confidence.
The built-in assistant translates AI findings into clear clinical language, flags signs of concern like stroke or tumor patterns, and suggests next steps—empowering doctors to act with clarity and confidence.
This tool visualizes changes across scans, detecting growth, shrinkage, or improvement in key brain areas. It supports ongoing monitoring and informed treatment adjustments over time.
This tool visualizes changes across scans, detecting growth, shrinkage, or improvement in key brain areas. It supports ongoing monitoring and informed treatment adjustments over time.
KEY OBSERVATIONS FROM THE EXISTING UI
KEY OBSERVATIONS FROM THE EXISTING UI
To help radiologists adopt AI confidently, the design focused on clinical clarity, control, and ease of use. AI outputs were made easy to interpret through structured summaries, color-coded overlays, and confidence levels. The interface was adapted to fit real diagnostic workflows—allowing radiologists to accept, reject, or prompt AI suggestions without friction.
To help radiologists adopt AI confidently, the design focused on clinical clarity, control, and ease of use. AI outputs were made easy to interpret through structured summaries, color-coded overlays, and confidence levels. The interface was adapted to fit real diagnostic workflows—allowing radiologists to accept, reject, or prompt AI suggestions without friction.
DESIGN APPROACH AND ITERATIONS
GleamerAI
GleamerAI
GleamerAI
MedicAI
MedicAI
Faster Diagnoses:
Clear visuals and structured insights speed up decision-making
Faster Diagnoses:
Clear visuals and structured insights speed up decision-making
Export buttons and controls were buried
Export buttons and controls were buried
Only one image view limited comparison
Only one image view limited comparison
Tools and filters were scattered
Tools and filters were scattered
AI diagnosis data lacked visual clarity
AI diagnosis data lacked visual clarity
Patient details were hard to locate
Patient details were hard to locate
Grouped Functionalities via a unified side panel for cleaner navigation
Grouped Functionalities via a unified side panel for cleaner navigation
Prominent Patient Context (age, sex, condition, ID)
Prominent Patient Context (age, sex, condition, ID)
Decision Controls to preserve clinical autonomy
Decision Controls to preserve clinical autonomy
Hover Labels to reveal region-specific AI insights
Hover Labels to reveal region-specific AI insights
Dual Image Views for comparing raw vs AI-annotated scans
Dual Image Views for comparing raw vs AI-annotated scans
Greater Trust in AI: Transparent, region-specific findings boost confidence and adoption.
Greater Trust in AI: Transparent, region-specific findings boost confidence and adoption.
Improved Patient Tracking: Scan comparisons support ongoing care and treatment planning.
Improved Patient Tracking: Scan comparisons support ongoing care and treatment planning.
Easy Onboarding: Intuitive design ensures quick learning and smoother clinical rollout.
Easy Onboarding: Intuitive design ensures quick learning and smoother clinical rollout.
Translating AI Outputs into Clinical Language
Learned how to bridge the gap between complex AI data and human-friendly, actionable insights that medical professionals can trust and act on.
Translating AI Outputs into Clinical Language
Learned how to bridge the gap between complex AI data and human-friendly, actionable insights that medical professionals can trust and act on.
Designing for High-Stakes Environments
Understood the importance of reducing cognitive overload and ensuring visual clarity when designing for fast, life-impacting clinical decisions.
Designing for High-Stakes Environments
Understood the importance of reducing cognitive overload and ensuring visual clarity when designing for fast, life-impacting clinical decisions.
Workflow-First Design Thinking
Focused on designing around the radiologist’s diagnostic flow — not just the interface — to improve usability and integration into real-world hospital settings.
Workflow-First Design Thinking
Focused on designing around the radiologist’s diagnostic flow — not just the interface — to improve usability and integration into real-world hospital settings.
Building Trust Through Interactivity
Realized that giving users the ability to accept/reject AI flags fosters transparency, builds trust, and strengthens adoption of AI-driven tools.
Building Trust Through Interactivity
Realized that giving users the ability to accept/reject AI flags fosters transparency, builds trust, and strengthens adoption of AI-driven tools.
To transform BrainSightAI's fMRI diagnostics into a transparent, clinician-first experience that accelerates stroke and tumor detection while ensuring radiologist oversight and trust.
iMerit
Radmat AI
Radmat AI
Existing Brainsight AI Platform
Low-fidelity wireframes
Redesigned Platform
Redesigned Platform
