Orchestrating the migration of Michigan Medicine's fragmented, Legacy system into a single, user-centered department intranet.

Role:


Lead UX Designer

Duration:


SEPT - MAy 2025

8 months

Tools:


Figma

FigJam

MS Sharepoint


Skills:


ui/ux design
INteraction design

Prototyping

Visual Identity

Agenda:

Overview

4 Problems

Before

Outcome

TImeline

REsearch

Growth Strategies

Clarity & Accessibility

Engagement & Story

Brand Identity

Impact

Next Steps & REflect

TL;DR

The approach

As a UX designer, I drove an intranet user-centered design, focusing on uncovering user mental models through research (surveys, card sorting, usability tests, focus groups) to build a new, intuitive Information Architecture from the ground up.

THe outcome

A unified and easy-to-navigate intranet built on SharePoint. The project's success was defined by our meticulous handoff, including design templates and documentation that ensured the client could easily manage and scale the new system independently.

1. Discover

2. Define

3. Design + Build

4. Handoff

Admin alignment

Scoping the problem

Generative research (surveys & focus groups)

Legacy system audit & Usability tests

Card sorting for site architecture

Synthesizing findings

User personas, scenarios & journeys

Insights to design requirements

MoSCoW prioritization

Information architecture design

Building live pages & templates in SharePoint

Iterative usability testing on live drafts

Refining UI & components based on evaluation testing

Deploying SharePoint pages/templates

UX specification

Final user flows & Interaction maps

Delivering adoption materials

Handoff presentations

Overview of the Phases of my design process for this project

MICHIGAN MEDICINE FAMILY MEDICINE COntext

MICHIGAN Medicine COntext

As a nationally recognized leader in patient care, education, and research, Michigan Medicine's Family Medicine department serves over 260,000 outpatients, housing 150+ faculty members, and supporting 39 residency programs. As the department scales, its intranet was intended to become the central, HIPAA-compliant hub for its community

The Problem

But the initial intranet migration had stalled. It failed to offer a compelling alternative to existing fragmented workflows, leading to chronically low adoption. Staff reverted to familiar, siloed tools, perpetuating disorganization and security risks. The tool that was meant to be an integrated community had instead led to wasted time, frustration, and reliance on ad-hoc, p2p requests, straining the entire system.

Our role

Our team was brought in with a focused: to improve the intranet's usability. However, our discovery research revealed a deeper truth: that usability fixes would not be enough; the core problem was a fundamental mismatch between the site's structure and the users' actual needs.

Our role evolved from designers to strategic partners. We acted as a bridge, connecting institutional security needs, departmental migration goals, and the workflow requirements of its users.

THe SOlutioN

We diagnosed the core issue as a failed information architecture and redesigned the entire intranet from the ground up based on user research. Our team then built a new system of role-based pages and templates directly within SharePoint, enabling real user testing and technical feasibility. Our final deliverable was a complete enablement package with detailed documentation, empowering the Michigan Medicine team to manage and scale the new intranet independently.

Final Design

Signifiers & CLarity

Site Architecture

Consistency & Templates

ADditional Changes

Final Design

Signifiers & CLarity

Site Architecture

Consistency & Templates

ADditional Changes

phase 1

DISCOVER

Understanding full context

Building Alignment — ADmin and Users

Why is an Intranet important?

An intranet is like your department’s internal website — a private space where only faculty, staff, residents, and team members can access and share information.

Think of this as a digital bulletin board, filing cabinet, and team hub rolled into one.

In a fast-paced, multi-site department like Family Medicine, an intranet centralizes clinical, educational, research, and administrative resources. An intranet helps by:

  • Centralizing important resources (forms, protocols, schedules)

  • Reducing email clutter and confusion

  • Making it easier to find what you need, when you need it

A well designed intranet supports day-to-day workflows, keeps everyone aligned, and helps you do your job, whether you’re a clinician, resident, admin, or researcher.

Initial Design

Home

Education

Administration

Research

Use Cases

The initial intranet was functionally abandoned. Users faced poor navigation, outdated content, and a search so unreliable they used public search engines for internal documents. Usage was driven by administrative mandate, not genuine need, offering no compelling reason for users to abandon their workflows.

🩺 Clinical Support

  • Quick access to OB precepting materials, clinical guidelines, or call schedules

  • Links to Epic help pages or templates

  • A one-stop shop for referral workflows, contact lists, or telehealth info

🧪 Research Coordination

  • Guidance on submitting grant proposals, IRB protocols, and departmental processes

  • Track research recognition, publications, or funding opportunities

🗣️ Communication & Culture

  • Internal announcements, department updates, and meeting minutes

  • Celebrations, retirements, awards, or new hires

  • Let people feel connected across clinics, academic divisions, and admin

📚 Education & Training

  • Residents can find their rotation schedules, didactic materials, or procedural logs

  • Attendings can post onboarding resources, policies, or evaluation forms

  • A shared place for M+Care curriculum content or faculty development workshops

📁 Administrative Efficiency

  • Find HR forms, PTO policies, or internal directories

  • Submit conference leave requests or reimbursements

  • Avoid emailing 3 different admins just to get a travel form

These are just some of the many use cases we documented throughout our research.

how might we?

How might we turn the intranet from an obligation into a go-to tool that supports and streamlines daily work?

Scoping the Problem

Project Objectives and GOals

Our engagement with Michigan Medicine began with a clear, but limited, set of client-defined goals focused on surface-level improvements to the existing intranet.

Initial Project Goals (Hover to see FInal Goals)

  1. Improve Usability and Drive Adoption
    Redesign the intranet to align with user workflows and reduce reliance on email.

  1. Strengthen Information Architecture
    Restructure content around departmental missions and ensure compliance.

  1. Ground Improvements in Research
    Use stakeholder interviews and benchmarking to inform improvements.

Initial Project Goals (Hover to see FInal Goals)

  1. Improve Usability and Drive Adoption
    Redesign the intranet to align with user workflows and reduce reliance on email.

  1. Strengthen Information Architecture
    Restructure content around departmental missions and ensure compliance.

  1. Ground Improvements in Research
    Use stakeholder interviews and benchmarking to inform improvements.

Challenge #1

During discovery, we reached a critical turning point. Scoping the project, even with two PMs, was a trial-and-error process. We were initially confident we could solve their intranet design problems, but our user research and workflow analysis revealed a deeper, more fundamental issue: this wasn't just a design problem; it was an information problem.

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

Applying best practices to the existing structure would have been like painting over a crumbling foundation. We shared our findings with the client, and together, we shifted our approach to address the root cause

Uncovering the real problem

Research

We set 3 research goals to ground our design process in user evidence and ensure we solved the right problems.

  1. Diagnose the intranet's current state to establish a baseline of user sentiment and identify key barriers to adoption

  2. Map existing user workflows to pinpoint specific sources of friction and prioritize areas for improvement.

  3. Determine core user needs and mental models that drive their habits and behaviors.

Audience

Users constantly dealt with high-stake workflows steeped in medical jargon and multi-personnel structures, so we prioritized understanding their specific processes and domain to design a scaleable solution.

Success METRICS

A Success measurement plan & our impact

Working with multiple PMs, scoping extent of the solution was a crucial part for the success of this project. We wanted to create an ambitious information-rich intranet that could be easily adopted and scaled. So some KPI's that defined as success were:

more Exposure & ADoption

Reach 75% Weekly Active Users (WAU)

Increased adoption rates across all mission areas and user groups.

more usability & enjoyment

Reach 90% User Success in terms of Completion Rate

Achieve 2x Faster Document Retrieval Across Teams

More Integration &
Collaboration

Enhance departmental collaboration through platform features.

Better alignment with HIPAA security compliance standards.

It was important for us to measure the success of our efforts and collaborate with the PMs and team executing the migration to provide a clear vision for the project’s goals.

Generative research

4 Methodologies

  1. Survey

  2. USability Test

  3. Card Sort

  4. Focus Group

To validate our initial hypothesis and understand the user's reality, we conducted a multi-method research plan. This deep dive was critical; it provided the evidence that shifted our project's focus from minor fixes to a fundamental redesign and allowed us to build our solution on a foundation of real user data.

phase 1: Quantitive Survey

THe Stakes… Painting a picture

n = 74 respondents

Our survey not only allowed us to quantify current usage, challenges, and awareness across the department, but also revealed a tool that had fundamentally failed to integrate into the department's culture and workflows.

Quotes:

"I never use it [...] honestly, I don’t even know what it’s good for.”

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

"I’m a lead for a few things on the intranet, but I can’t update my own page [...] It’s out of date, and emailing changes is exhausting."

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

"I regularly need to know who the medical directors are [...] but that info is hard to find or outdated on the intranet.”

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

Key Findings:

  1. Functionally Abandoned: 69% of respondents never (29%) or rarely (40%) used the site; 28% didn’t know it existed.

  2. Email & Workarounds Dominated: Email was the primary communication tool (35%); users preferred Dropbox and Google Drive for collaboration.

  3. Need for Reorganization: Users wanted clearer organization and easier access to resources. One clinician noted, “The site is heavily oriented to clinical faculty… I cannot find information that is helpful for research processes.”

Significance:

  1. The Bar for Success is High: The solution must be significantly better than existing tools to win users back.

  2. Information Architecture is the Central Problem: The core issue is not about colors or fonts, but about a fundamentally broken organizational structure.

  3. Personalization is Non-Negotiable: A role-based approach is the only way to make the intranet relevant to its diverse audience.

phase 2: Usability Tests and System audit

phase 2: Usability Tests & System audit

system Audit and Usability test

n = 7 Participants

For phase 2, we conducted an audit to fully flush out the initial intranet, mapping out information architecture, checking for accessibility, and noting errors. Information discovery, discovering redundancies and

Then to further understand the user journey, we explored task flows, devised tasks, prioritized them, and conducted 1 on 1s, meeting our users where it was most convenient for them, identifying more nuanced pain points in navigation and user behaviors for our audience. (What's a better way to phrase in consideration of our users we went to their office to conduct tests, giving our project exposure, building stakeholder alignment, giving us a better opportunity to get our users environment and workflow.)

Key Findings:

  1. The Nav Bar Was the Primary Mental Model: All users immediately looked to the main navigation bar to begin a task, expressing frustration when forced to scroll or hunt for information on the page itself.

  2. Search Was Broken and Untrustworthy: The search function was so ineffective that users refused to engage with it, with one stating they were "not gonna go through three pages of random file[s].

  3. The IA Was "Clunky" and "Unfamiliar": Users consistently struggled with the content organization, describing it as illogical and reporting difficulty finding information even when they knew it existed.

phase 3: Card Sort

Card Sorting Activity

n = 4 respondents

Of phase 3, we strived to uncover how information was organized within our users, then how they would group and label content based on their intuition, setting a great foundation for our Information Architecture.

Key Findings:

  1. Users Organize by Role & Mission: Participants grouped content by mission area (e.g., Clinical, Research) or role (e.g., Residency, Faculty), revealing that the current structure doesn’t match user mental models.

  2. “Quick Links” = High-Frequency Tools: Users saw “Quick Links” as a primary navigation space for essential tasks—especially Documents & Forms and Logos & Branding—not just a list of miscellaneous links.

  3. People = Critical Resource: Users repeatedly emphasized the need for colleague directories and contact details tied to specific responsibilities—highlighting people as a core resource.

phase 4: Focus Group

FOcus Group

n = 23 Participants

Our focus group focused on hearing direct feedback with a diverse set of members from stakeholders to residents, uncovering cultural blockers and defining the intranet. This method was extremely effective in building empathy and importance around our project for them, materializing our impact.

Quotes:

"With our schedules, we will not self-study. It needs to be immediately obvious."

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

"It’s not just about documents; we need to know who to talk to [...] and who’s in charge of keeping each SharePoint page updated."

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

"The intranet should be our homepage for day-to-day operations [...] the first place we go."

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

"What makes Google Docs so useful is the fluid sharing [...] That’s the level of collaboration we’re used to."

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

Surprise!!

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

One member of the focus group was an individual of low-vision and was vocal about how important it was for the intranet to be screen reader accessible. This was a small detail and surprise interaction that shaped our delivery!

Scoping this project, even with two PMs, was a process of trial and error. At first, we were confident we could solve their intranet design problems, but user research and workflow feedback revealed our limiting factor was an information problem. We shared these findings with the team and adjusted our approach accordingly.

Key Findings:

  1. Learning is a "Daunting" Task: Participants found learning the intranet daunting and preferred departmental training over self-study. This mandated an intuitive, zero-learning-curve design.

  2. Clarity on "Who Does What": Users requested clear information on colleague responsibilities to know who to contact for specific topics, validating the need for an enhanced directory.

phase 2

Define

Transforming Data to Action

FEedback -> Insights

Synthesis the "Aha" Moment!

Our affinity map synthesized hundreds of data points into a clear verdict: minor fixes would not work.

This synthesis was our point of realization, providing the undeniable evidence that our initial scope was flawed and a simple redesign would fail.

Evolving the Approach

Project Objectives and GOals (REfined)

We shifted from workflow fixes to an integrative redesign and with a new understanding of actual user requirements and a targeted project scope.

Final Project Goals (Hover to compare)

  1. Target High-Impact Improvements
    Map existing workflows to uncover hidden pain points and understand what’s preventing effective use of the intranet today.

  1. Define the Ideal User Experience
    Develop a new information architecture and toolset that reflect user mental models and reduce friction in task completion.

  1. Enable Seamless Collaboration
    Bridge communication gaps across departments. Design the intranet to work with—not against—existing platforms and tools critical to staff workflows.

  1. Optimize for Adoption
    Enhance the SharePoint implementation with proven best practices and design a seamless onboarding experience to drive user adoption. Identify key communication barriers and design features to streamline interdepartmental workflows.

Final Project Goals (Hover to compare)

  1. Target High-Impact Improvements
    Map existing workflows to uncover hidden pain points and understand what’s preventing effective use of the intranet today.

  1. Define the Ideal User Experience
    Develop a new information architecture and toolset that reflect user mental models and reduce friction in task completion.

  1. Enable Seamless Collaboration
    Bridge communication gaps across departments. Design the intranet to work with—not against—existing platforms and tools critical to staff workflows.

  1. Optimize for Adoption
    Enhance the SharePoint implementation with proven best practices and design a seamless onboarding experience to drive user adoption. Identify key communication barriers and design features to streamline interdepartmental workflows.

4-Step Definition

We integrated raw feedback into insights and a clear design foundation through a 4-step definition process.

Defined our users through personas, their context through scenarios, their emotional journey through the journey map, and our priorities through design requirements.

  1. Personas

  1. User Scenarios

  1. Journey Maps

  1. Design Requirements

  1. Personas

  1. User Scenarios

  1. Journey Maps

  1. Design Requirements

Research Wrap up Presentation

What did we discover about our users and their problem space?

We presented a comprehensive summary about our research findings to our stakeholders, management team, and PMs. This way we were able to put context behind our requests and initiatives and guide with clarity for efficiency.

phase 3

Design and Build

Broad Ideation -> FOcused, pragmatic Implementation

Defining Possibilities

IDEAtion and Sketches

We independently sketched a wide range of solutions—from interactive dashboards to novel navigation patterns—before converging on the approaches aligned best with out design requirements.

Embracing Real world constraints

Skipping wireframes to Implementing in Sharepoint

Our next step was a critical, pragmatic pivot. Instead of creating wireframes, we began building our mid-fidelity designs as live draft pages directly within SharePoint. This hands-on approach was a strategic decision that ensured every design concept was immediately tested against the platform's real-world constraints.

Example #1

This shows the two different card options from the draft of the Residency Program Teaching Opportunities Page (later the Residency Page).

We debated two primary approaches: a general announcement-style card versus a detailed event information card.

These two screenshots were our mid-fidelity designs for our Document ad Forms and About Us Page.

We experimented with different layouts and ways to establish information hierarchy, navigation clarity, and scannability. At this stage, the focus is on layout, structure, and content prioritization.

Example #2

Example #3

The first iteration of the Research page was designed to chunk large amounts of information into digestible, navigable sections and make the page more approachable for users.

Early feedback suggested that the original Research resources felt overwhelming and difficult to browse, so our design intent was to reduce cognitive load and improve findability.

Top-Down Approach

Deconstructing the Old to Build the New

With our build environment established, we deconstructed the failed intranet and rebuilt the Information Architecture using a "top-down" strategy. Why top-down? By organizing the entire site around the user's primary mental model—their mission area—we established clear, predictable "memory anchors." This approach creates natural chunking consistent with how concepts are mentally organized, allowing users to quickly grasp where they need to go without conscious effort.

Improving Engagement and Storytelling

To address user anxieties and career goals, we created urgency and brought Soojin's UVP to the forefront.

This meant the structure and storytelling of the homepage was essential to effectively market her brand. Research from our competitive analysis, 4 iterations, user testing, and 3 client meetings depicted a storytelling formula to most effectively convey Soojin's services and compel clients.

The structure was : CTA, Stakes, Client network(Credibility), Value Proposition, Metrics(Proof), Plan, Testimonials.

Sketch

———>

Wireframe

———>

High-fidelity

High-Fidelity

To embed a website or widget, add it to the properties panel.

Browse through our final design!

Brand Identity Guidelines

A brand is more than just a logo and visual identity. It also represents your recognition & perception, trust & credibility, ability to market, and serves as the connection between you and your customers/clients.

This is why we created a brand identity guideline to act as a foundation for her website, presentations, seminars, online courses, workshops, etc.

To embed a website or widget, add it to the properties panel.

Click to go through Brand Guidelines

Impact

(Metrics 2 months after implementation)

41%

Increase in unique visitors

26%

Decrease in bounce rate

300+

New followers on LinkedIn

Brand clarity and a foundation for sustainable growth

🎊 Following our consultation, Soojin has successfully 🎉

Soojin has followed through with some of our recommendations!

  • Participated in 2 collaborative seminars, expanding her professional network and client reach

  • Obtained ICF certification, boosting her professional credibility and ability to command premium rates

  • Created online modules through prestigious partnerships with University of Michigan College of Engineering and Emeritus

  • Regularly posts professional insights and personalized stories on LinkedIn, building her brand presence

  • Several other recommendations have been implemented but details remain confidential

  • A version of our design is current live! Client might've made some tweaks. talentpeakadvisors.com

Next Steps and Reflection

I was too locked into taking this photo T-T Team <3

We presented all our research and reports to Soojin as well as design changes and growth actionables. The released version of the website has been well received by our client, and launched on SquareSpace. However, there are still be improvements to be made, but now Soojin should be able to filter her data even more, and customize her brand based on its needs.  

We helped revised a foundation that allows our client to build on top of, dissecting 5 industry services, and creating artifacts for growth. The patterns and designs created would later help expand Soojin's brand, network, and future products as we constantly strive to improve the TalentPeak Experience.

Reflection

This experience working with Soojin and my team has been invaluable. During the initial months, meetings with her were challenging due to rigorous professional standards. She set a level of excellence for her and her team. After 3 discouraging meetings that left us overwhelmed and her presence overbearing, I realized we didn't establish expectations and mutual trust. In essence she wasn't convinced we were capable of working with her. With this realization I was able to lead a meeting that boosted our momentum for future interactions, gaining her respect and nurtured a better team experience.

Thanks for stopping by! Look forward to chatting ;)

NAVIGATION

HOME

SELF

Case Studies

UM MEdicine

Talentpeak

uNFOLD

NoAA

Umich CAmpus

Locali

Get in TOuch

LINKEDIN

EMAIL

INSTAGRAM

Extras

RESUME

THOUGHTS

KEVXU@UMICH.EDU

Designed & made by @ KEVXU 2024

UPDATED 09.11.2025

HOME

Self

in

cv