Portfolio · Ashleigh Hearn · Content Design
A before and after content transformation project demonstrating modular design, HTML authoring, and AI assisted workflows
Project goal: Identify a gap in Uber's existing rider support content, redesign it into a modular agent ready format, and hand code it in HTML following content architecture and design system standards.
When I researched how Uber currently handles support content around driver cancellations, I found that the information was spread across three separate help articles. There was no single resource that walked a rider or support agent through the full picture in one place. Each article covered a piece of the topic but none of them connected the dots in a way that made the resolution process clear and fast.
On top of that the existing content mixed rider cancellation policies and driver cancellation policies together in the same paragraphs with no visual separation or structure. A support agent trying to help a rider who was charged after their driver canceled would have to read through the entire article to find the relevant policy. That slows down resolution time and increases the chance of giving the wrong information.
Before: Example of the original Uber content
"You can cancel your ride at any time through the Uber app, but may be charged a cancellation fee if you cancel after you are matched with your driver. Cancellation fees pay drivers for the time and effort they spend getting to your location. For most economy ride options if rider canceled 2 or more minutes after requesting. For most Premium ride options if rider canceled 5 or more minutes after requesting. If your driver cancels a trip after they have been waiting at your pickup location for a specific amount of time, a cancellation fee will apply. These wait times vary depending on which Uber product you have requested. If you believe you were charged a cancellation fee in error, let us know."
Issues identified: No headings. Rider and driver policies mixed together. No decision tree. No next steps. Difficult for agents to use quickly.
I redesigned the content into a single structured HTML page organized around how a support agent actually uses the information. The page starts with what happens immediately after a driver cancels, moves into a decision tree that covers the three main fee scenarios a rider might face, and ends with clear numbered steps for disputing a charge. Every section is focused on one thing so an agent can get to the right answer without reading content that does not apply to their situation.
I started by reviewing the three existing Uber help articles to understand what information existed and where the gaps were. I then used Claude as an AI writing assistant to help me outline the content architecture and identify the key decision points a rider or agent would need to navigate. From there I wrote and structured the content myself and hand coded the full page in HTML using inline styles throughout.
I built a color coded decision tree to separate three fee scenarios so agents can immediately see which situation applies to their customer. I structured the ride type wait times into a reference table for quick scanning instead of burying them in paragraph text. I also applied metadata tags including content type, topic tags, audience, and last updated date to support search optimization and AI discoverability.
The goal throughout was to create something an agent could open mid conversation with a customer and use immediately without having to search for the relevant section or cross reference another article.
| Job Requirement | How This Project Shows It |
|---|---|
| Transform legacy content into modular formats | Restructured dense unorganized content pulled from three separate articles into clearly separated modular sections |
| Hand coded HTML with inline styles | Entire page built in HTML using inline styles throughout with no external stylesheets |
| Apply metadata for search and AI discoverability | Meta tags and a metadata footer added with topic tags, audience, and content type |
| Build decision trees and step by step guides | Color coded decision tree and numbered dispute steps built directly into the article |
| AI assisted content workflows | Used Claude to outline content architecture and identify decision points before writing |
| Conduct quality audits | Identified structural and accuracy issues across three original articles before redesigning |
The content used in this project was sourced directly from the following Uber help articles:
Ready to see the finished article?
View the Redesigned ArticleContent Type: Case Study | Audience: Hiring Managers and Content Teams | Topic Tags: content design, HTML, modular content, knowledge management, AI assisted writing | Last Updated: May 2026 | Owner: Ashleigh Hearn