Project details
Client:
Orevan
Tool:
Figma
Team: Moamen Ali - Youssef Haider - Perihan Medhat - Youssef Ahmed - Mohamed Adam
OrviAI – AI-Powered Medical Presentation Assistant
Project Summary
OrviAI is an intelligent assistant designed to streamline how healthcare professionals (HCPs) search, review, and assemble scientific training content from internal PDFs and PowerPoint files. The platform simplifies access to complex data, enabling faster and more accurate presentation preparation — with built-in citations and image sourcing to maintain credibility.
As the UX designer, I was responsible for translating complex AI and data processing tasks into a seamless, intuitive experience tailored to the workflows of HCPs and admins.
Design Problem
HCPs spend significant time manually navigating dense presentations and research files, often resulting in:
Inefficient preparation of training materials
Missed visual references (charts, diagrams)
Difficulty tracking sources and citations
The design challenge was to make the interaction with AI-generated content feel trustworthy, transparent, and human-centric, especially in a medical setting where accuracy is non-negotiable.
UX Solutions & Key Features
🎯 Natural Query Interface
A clean, search-driven interface allows users to ask questions in everyday language — eliminating the need to browse or scan long documents. Answers are clearly structured with references to slides or pages.🖼️ Text + Image Response Cards
Responses are displayed as smart cards showing:Answer
Source (e.g., "Slide 7, Diabetes.pptx")
Image thumbnail (if available)
Users can click to preview the full document context — bridging AI output with original material.
📁 Content Explorer with Filters
Enables browsing and filtering by:Material type (PDF/PPTX)
Category (e.g., Oncology, Cardiology)
Source company
Date added
This gives users more control over the data they rely on.
🧠 Answer History & Favorites
HCPs can revisit past queries, save useful results, and export answer cards for use in their own presentations or training sessions.🔒 Trust-Centric Design
Focus on source transparency, inline citations, and preview functionality to build trust in the AI system. Every answer is clearly attributed to its origin with no "black-box" moments.🛠️ Admin Dashboard (For Material Curation & Analytics)
Designed a back-office tool for admins to:Upload and tag new files
View platform usage and popular queries
Manage companies, categories, and users
The interface uses a dashboard-style layout with clean tables, filters, and graphs to give admins control and visibility.
UX Process
Conducted stakeholder interviews to map the real-world training preparation workflow.
Created persona profiles for:
HCPs (end users)
Medical content managers (admins)
Designed low-fidelity wireframes → validated with mock user flows
Developed interactive prototypes and design systems with a focus on clarity, accessibility, and responsiveness
Worked closely with AI engineers to ensure a balance between complex system outputs and user-friendly presentation
Impact
By embedding AI into a user-friendly, searchable, and transparent experience, OrviAI:
Reduced average presentation prep time by 40%
Improved accuracy by ensuring every claim is evidence-linked
Increased platform trust with visual traceability and clean UX