HarshithaKommaraju

UMD DEPT OF TRANSPOR-TATION

WEBSITE REDESIGN

Redesigning the complex charter request flow to support clarity, speed, and user independence

ROLE

UX Design & Research, Team of 4

TIMELINE

24 hour make-a-thon

CLIENT

University of Maryland, Dept of Transportation

OUTCOME

First place winner

CONTEXT

UMD’s charter service allows faculty, staff, and coaches to request transportation for classes, events, and athletic travel. The existing experience relied on a dense web form and scattered information, creating confusion, errors, and frequent dependency on staff support. This redesign focused on improving clarity, streamlining the request process, and enabling users to complete submissions independently.

CHALLENGE

How might we redesign the UMD Charter Service request experience to improve clarity, streamline the process, and promote user independence while submitting a request?

The solution needed to work within:

UMD’s existing design system

Institutional and legal constraints

Limited access to real users

A 24 hour design timeline

MY CONTRIBUTION

While this was a collaborative team effort. my primary contributions included:

Leading research synthesis and problem framing

Developing the sequence model to understand the end to end booking journey

Conducting competitor analysis to identify usable patterns

Designing the vehicle selection model and scannable vehicle cards

The design and logic of “Best Fit” recommendation feature

UNDERSTANDING USERS

We grounded decisions in insights from a faculty member who had previously used the service, feedback from UMD transportation staff based on recurring complaints and heuristic evaluation of the existing site and form

1. The Request process lacked structure

Users had no sense of progression or completion, leading to mistakes and abandoned submissions.

2. Dense, unstructured in puts overwhelmed users

Large blocks of fields without hierarchy caused hesitation and missed requirements.

3. Text heavy layouts and inconsistent patterns reduced confidence

Users struggled to scan information and felt uncertain about their choices

THE REDESIGN.

LONG SINGLE FORM → GUIDED REQUEST FLOW

The original one page form was restructured into a sectioned, step by step experience with clear progress indicators and a defined path to completion.

This reduced cognitive load and helped users understand where they were and what remained.

The PDF version prior to redesign

A visual guide on the charter request process at the top of the page

The pdf converted into a web form with logical steps and form validation

The progress indicators to guide users while filling out the form

VEHICLE SELECTION DESIGNED FOR SPEED

OVER EXPLORATION

Design Tension

Early concepts explored an interactive, exploratory vehicle gallery. Persona refinement revealed that users prioritized speed and certainty over browsing, leading to a simpler, decision focused layout.

Vehicle information previously presented as bullet points was redesigned into scannable cards that surface key specifications first, with expandable details for deeper inspection.

A “Best Fit” recommendation is calculated using passenger count, airport travel, and ADA needs.

Vehicle rates, availability and

info spread across multiple

access points

All relevant info gathered in one card

Collapsible rates to save space and provide relevant info when needed

Responsive design as a focus: Mobile view of the vehicle info

“Best Fit” and “Not a Fit” generated based on users passenger and trip requirements

PROGRESSIVE DISCLOSURE, ADDED CONTEXT, AND

SMART DEFAULTS

Inputs were reorganized to reveal information only when relevant, with clearer distinctions between required and optional fields.

Autofill and sensible defaults reduced effort and hesitation.

Unclear and large amount of information in the middle of the form

Additional context for all the fields with next steps laid out

Concise view of vehicle info within form to facilitate decision making

Context based adaptive input fields to lower cognitive load

PREVENTING ERRORS THROUGH VALIDATION

Pickup and drop off fields were redesigned with interactive location selection and format validation, reducing ambiguity and follow up clarification.

Input validation with Google API connectivity for locations

RESULT

In 24 hours, we delivered a fully clickable, responsive hi-fi prototype that:

Clarified complex character terminology

Streamlined the request process end to end

Reduced reliance on staff support

Worked entirely within UMD’s design system

The design and logic of “Best Fit” recommendation feature

REFLECTION

This project reinforced the value of restraint under pressure. Rather than adding features, we focused on structure, hierarchy, and decision clarity. Working within real institutional constraints sharpened our ability to design solutions that were both user centered and implementable.

V-CONNECT COURSE DELIVERY PLATFORM REDESIGN

ROLE

UX Design + Research

TIMELINE

3 months

STATUS

Product Launched

THE QUICK VIEW

STARTED AS

Complicated Internal B2B Tool

Research

Diagnosed that the problem was structural, and not visual

Design

Redesigned user workflows to be simpler and more efficient

Iteration

Incorporated feedback from usability testing and stakeholders

Dev Collab

Continuously worked with development team till product launch

SHIPPED AS

Live product now with 30+ corporate clients

PROBLEM

V-Connect was Potentia's internal course delivery platform, built for their own instructors and never meant for outsiders. When Potentia decided to sell it as a standalone product, the problem became impossible to ignore.

The platform only made sense to the people who had already spent years inside it.

New users called back constantly. Even client HR managers who had been granted partial access needed hand-holding through basic tasks.

Veteran facilitators had built mental models over years of repetition that made the system feel natural to them. For anyone coming in fresh, those same jumps were walls.

INSIGHTS

I focused on existing internal facilitators, aka the people who knew the system best. My research questions focused on identifying where the friction actually lives, and what invisible knowledge would a new user have no access to?

“Despite using this for years, I still get lost sometimes and need to find my way back. I basically just have a bunch of routes memorized that I use when I need to finish a task”

- Facilitator who had used the product for 3+ years

Three findings shaped everything that followed

BATCH CREATION WAS ERROR PRONE AND COGNITIVELY DEMANDING

No clear structure through the multi-step process. Users relied on verbal knowledge passed down. For every new user, even one wrong step meant starting over or calling someone.

LARGE UNSTRUCTURED FORMS OVERWHELMED USERS

No hierarchy between required and optional inputs. Users stalled, skipped critical fields, or guessed with no recovery path when they got it wrong.

DENSE, INCONSISTENT INTERFACE MADE SYSTEM NAVIGATION HARD

Long text-heavy tables with inconsistent visual patterns made it impossible to get an overview at a glance. Users couldn't find the information needed quickly

If we were selling this to people who had never touched it before,

no amount of visual cleanup was going to fix a broken process.

DESIGN DECISIONS AND WHY?

SCATTEREED PAGES

→ GUIDED FLOW

Before touching any individual screen, I mapped the entire task flow and rebuilt the sequence from scratch. Batch creation was the most cognitively demanding part of the product. Facilitators navigated between disconnected pages with no sense of where they were or what came next. The fix seemed obvious: a linear step-by-step flow with clear progression indicators at every stage.

INITIAL IDEA

A user flow restructured to have a linear process:

Create Batch → Add activities → Add participants → Publish

PUSHBACK

Facilitators flagged that one of V-Connect's actual strengths was the ability to add activities and participants in parallel or in alternating order. Facilitators relied on this flexibility yet the linear flow removed that feature.

RESOLUTION

The step-by-step structure stayed, but the progression model changed. Instead of moving users forward automatically, the guided flow became a choice-based navigation process where facilitators choose what they want to work on next, and the following screens adapt based on that. First-time users get structure and clarity while users who know what they're doing keep their flexibility.

Optional shortcuts for experienced users

Progression indicators within each step

New Batch

Step-by-step guided flow where you choose your own path, instead of a fixed sequence

ITERATION 2

REDUCING COGNITIVE LOAD IN FORMS

Some of the most-used forms in V-Connect had upwards of fifteen fields on a single screen, with no distinction between what was required and what wasn't.

I broke any form above a certain input threshold into staged steps.

INITIAL IDEA

I broke forms into staged steps and updated internal jargon to be more explicit, adding contextual help throughout, visible all the time

PUSHBACK

Stakeholders felt the help text was clutter and unnecessary for anyone who knew the system, taking up valuable screen space.

RESOLUTION

I kept the clearer language but made help text collapsible and tooltips hover-triggered. Users who need context get it instantly. Users who don't can ignore it entirely. Neither group compromises.

Required vs Mandatory made clear

Collapsible functionality to reduce overwhelm

Suggested defaults reduce decision fatigue

Contextual Guidance on hover

Discussion thread with public

posts and individual replies

MAKING INFORMATION DIGESTIBLE

The batch list cards were carrying too much at once including the status indicators, progress metrics, facilitator roles, and action buttons all competing for the same space. Clickable and non-clickable elements lived side by side with no clear distinction between them.

INITIAL IDEA

I converted the text-based info into icons and added a divider between informational icons and action icons to create separation.

FEEDBACK FROM USABILITY TESTING

It was still taking too much space, and the team pointed out that not all that information was needed on first glance since users scanning their batch list were looking for specific things, not everything at once.

RESOLUTION

I stripped the cards back to only the information users actually needed immediately, paired with the action icons. Everything else like the status of publishing, scheduling info, etc moved into the filtering and sorting system. Information you need when scanning lives on the card while information you need when you're looking for something specific lives in the filter. The cards got cleaner, and the filtering became genuinely useful rather than decorative.

Search, filter, and sorting for info not in cards

Horizontal tabs to make large data sets navigable

Data visualizations replacing text heavy indicators

Only essential info visible on cards with detailed info moved to filters

Filter

Sort

OUTCOME

V-Connect launched to market and is now live. 30+ client companies. 70+ course batches delivered. 1,800+ participants through the platform. A tool that once required an internal expert to navigate is now being used independently by people who had never seen it before.

30+

CLIENTS

1,800+

PARTICIPANTS

70+

COURSE BATCHES

REFLECTION

While the road to the final designs involved countless iterations and feedback sessions, the final product conveys the benefits of actively listening and incorporating feedback throughout the UX lifecycle. Redesigning the system and process instead of the visual interface reenforced the need to view products from a wider lens than only design and further intrigued me towards system based UX instead of visual based UX.

What shipped was meaningfully different from what I inherited.

PROBLEM

RESEARCH

DESIGN DECISIONS

OUTCOME

REFLECTION