Frank Pelosi III

Projects

About

Resume

Frank Pelosi III

Projects

About

Resume

Frank Pelosi III

2025

Medeligo

Reimagined oncology research for the web with integrated AI functionality.

Overview

Our team redesigned Medeligo, an iOS oncology research app, into a responsive web platform better tailored to oncologists’ workflows. We also explored AI-driven features to enhance support and clarify Medeligo’s market niche. Our solution is a flexible oncology research platform that streamlines document discovery and storage, saving clinicians time and adapting to their preferences.

location_pin

Location

Ann Arbor, MI

work

Role

Product Designer

group

Team

3 Product Designers

event

Timeline

Jan - Apr 2025

CONTEXT

Adapting and expanding an iOS native oncology research platform to the web with AI.

Medeligo is an iOS app that helps oncology professionals find vetted, up-to-date research to support clinical care. Our redesigned platform offers a flexible search experience—combining AI and traditional methods—and introduces essential document storage functionality, in order to better match oncologists’ workflows and preferences.

PROBLEM

Fragmented and inefficient oncology research delays care and reduces clinical effectiveness.

Clinical research for oncology professionals can be a time consuming and liable process due to the granular detail of information and high patient volume. Oncologists have minimal time to review research and challenges validating materials that are vetted and up to date.

Struggle to identify answers to key questions in granular material.

Face ambiguities in validating the latest oncology research.

Time constraints in providing efficient and accurate patient care.

Business Needs

As an iOS app, Medeligo (1) limits its user base and (2) lacks integration with desktop, tablet, and third-party systems like Electronic Medical Record (EMR) systems.

A web-based platform expands access and better aligns with clinician workflows. Stakeholders also sought to explore AI-driven features to streamline research and position Medeligo as a leader in AI-powered medical research.

As an iOS app, Medeligo (1) limits its user base and (2) lacks integration with desktop, tablet, and third-party systems like Electronic Medical Record (EMR) systems.

lightbulb

How Might We…

How might we design a flexible, web-based Medeligo platform that enables oncology professionals to access reliable, up-to-date research supports diverse clinician workflows?

RESEARCH

A wholistic approach to transforming oncology research.

Our research strategy combined qualitative and evaluative methods to build a well-rounded understanding of user needs and product context. We aggregated our findings in an affinity map to better inform ideation and project scoping.

We implemented the following methods:

User Interviews (5)
Competitive Audit
Interaction Mapping
User Flow Diagramming

In addition, we also leveraged existing research:

Existing Interviews (5)
Heuristics Evaluation

We implemented the following methods:

User Interviews (5)
Competitive Audit
Interaction Mapping
User Flow Diagramming

In addition, we also leveraged existing research:

Existing Interviews (5)
Heuristics Evaluation

Research Insights

alarm

Clinicians struggle to balance patient care and granular research within limited time.

From our research, clinicians only have around 5 minutes between patients to prepare cases and research as needed.

fact_check

Credibility of information used for diagnosis and learning is essential.

All interviewees prioritize peer reviewed journals as being the most credible. Followed by national databases.

folder

Many oncologists have no efficient way to save, sort, and share documents.

All oncologists interviewed use single folder system to save and sort, with minimal organization.

warning

The iOS experience has critical performance and information architecture issues.

Users cannot select multiple cancer subspecialties, types, or source types at once, creating poor experience.

User Personas

From our initial research and stakeholder feedback, we created (2) target user personas to represent both veteran and novice oncology professionals.

Dr. Jenna Knox

Age:

Occupation:

Oncologist

Experience:

23 Years

Goals:

Stay informed about latest clinical trials, therapies, and treatment protocols for domain.
Contribute to research in field.

Frustrations

Limited research in pediatric cancers.
Time consuming research.
Lack of platform integration.
Rapidly changing guidelines.

Miles Norris

Age:

Occupation:

Registered Nurse

Experience:

1 Year

Goals:

Enhance knowledge of oncology for better patient experiences.
Improve communication with oncology floor.

Frustrations

Balance patient care with research.
Information overload.
Challenge staying up to date with research.

IDEATION

Discovering ways to streamline research process and optimize organization.

Our team took a structured approach to ideating 3 unique concepts, which helped minimize scope creep and keep both user and business goals in mind. For each concept, we created (1) rough sketches of key pages (2) user storyboards (3) a site map, and (4) user flow diagram. The concepts are as follows:

Concept 1: Multi-Faceted Search

Concept 2: Collaborative Newsfeed

Concept 3: AI Case Manager

Concept Testing

Our team conducted 3 moderated user concepts tests using mid-fidelity designs of the 3 concepts aforementioned concepts, recruiting physicians from personal and professional networks.

Each interview lasted roughly 30 minutes and consisted of both open-ended questions and Likert scale rankings of each concept’s respective features.

Our team conducted 3 moderated user concepts tests using mid-fidelity designs of the 3 concepts aforementioned concepts, recruiting physicians from personal and professional networks.

Each interview lasted roughly 30 minutes and consisted of both open-ended questions and Likert scale rankings of each concept’s respective features.

Concept Testing Insights

target

AI-Powered search needs high accuracy and relevance.

input

Case optimization needs less input effort.

handshake

Collaboration and newsfeed features are less essential

question_mark

Clarifying questions & in-depth research are distinct user needs.

Bringing It All Together

From our concepts tests we agreed to combine Concept 1: Multi-faceted Search and Concept 2: Case Manager into a comprehensive solution.

The AI driven features in both of these concepts were the highest rated from our testing. Including a standard search is important given users slight mistrust of AI with highly technical and essential research.

Our team conducted 3 moderated user concepts tests using mid-fidelity designs of the 3 concepts aforementioned concepts, recruiting physicians from personal and professional networks.

Each interview lasted roughly 30 minutes and consisted of both open-ended questions and Likert scale rankings of each concept’s respective features.

Consolidated Features List

auto_fix

Conversational Search

An AI powered search allowing oncologists to ask clarifying questions using plain business English.

filter_alt

Standard Search

Advance drill-down capabilities for oncologists who are familiar with common research patterns or uncomfortable with AI.

mic

Live Sessions

This feature acts as AI assistant, helping physicians transcribe meetings, offer key insights, and provide recommendations.

bookmark

Case Manager

Provides a comprehensive way to organize resource and leverage additional AI for summarization and discover.

Responsive Mid-Fidelity Designs

Our team took a desktop-first approach, as no participants thus far indicated they conduct oncology research on a mobile device. We also considered 3rd party integration by including a collapsable aside panel.

HIPAA Complications & AI Overload

Stakeholders had concerns of HIPAA compliance of Case Manager Clinical Features and Live Sessions. Pursuing HIPAA compliance does not meet business goals. These features were removed in favor of a simplified document storage system. Stakeholders also favored a stripped down case manager; a simple way to access research materials.

We removed these features prior to designing V2.

Style Guide & Branding Assets

Typography

Color

Grid

Icons

Components

Brand Assets

DESIGN & ITERATION

Refining designs for user testing and iteration.

Given stakeholder feedback, our team refined our project scope and integrated our high fidelity design system prior to user testing. We notably (1) removed HIPAA violating product features and (2) Refined verbiage of features to improve clarity and (3) enhanced visual language.

Our final feature list is as follows:

auto_fix

AI Search

filter_alt

Advance Search

bookmark

Library Storage

User Testing Flow

User Testing

Our team conducted 3 semi-moderated usability tests with medical professionals consisting of 7 key tasks using an interactive prototype. Usability metrics collected included time to task, error number, and qualitative feedback via the think aloud method.

FORMAT_QUOTE

Clickable citations were expected in the conversational search citations.

100% of users expected to see citations upon query output.

tune

Standard search drill-down tasks were too complex

Credibility of information used for diagnosis and learning is essential.

66% of users struggled to complete standard search progressive disclosure.

Interest in more granular source filtering for clinical trials.

Many oncologists have no efficient way to save, sort, and share documents.

All users expressed an interesting in filtering clinical trials further, by phase.

chat

Add intelligent, contextual follow-up suggestions.

2 users wanted follow-up options to drill down AI search.

SOLUTION

A flexible, intelligent oncology research platform.

The new Medeligo web app reflects the future of clinical research as a flexible, integrated system designed to support oncology professionals find and manage in a fast-paced environment.

Key Features

auto_fix

AI Search

Allows users to ask questions using plain business English
Citations included in every response
Voice to text functionality to support oncologists on the go

bookmark

Library

Use custom or quick access folders to store research material.
Discovery tab to find resources based on user activity.
Intuitive quick process to generate new research folders.

filter_alt

Standard Search

Leverages familiar visual patterns to search and filter.
Drill down based on cancer subspecialty and tumor type.
Scalable filter system to include more granular source filters.

Notes on Technical Feasibility

Prior to finalizing our designs, our team consulted a senior MLP expert to discuss the technical feasibility of the new Medeligo platform. All finalized features were approved based on their technical feasibility, though questions still remain regarding data structure and storage implementation.

REFLECTION

Creating a scalable framework for AI-driven research.

This project strengthened my ability to balance user needs and business goals in an objective-oriented project, while also giving me the opportunity to push new product features and build a comprehensive go-to-market strategy. Our solution addresses the following areas: