Children's Hospital of Philadelphia: Module Discovery Tool

Gooey.AI is a low-code copilot builder enables impact organizations to create custom chatbots for support to farmers, nurses and other users.

When I joined the project, the copilot builder was primarily designed for technical users, creating significant barriers for the program managers who required weeks of support team assistance to successfully launch their first chatbot.

I worked with the Gooey.AI team to redesign the entire user experience, transforming the copilot builder into an intuitive, step-by-step process.

Timeline

April 2025 - June 2025

Role

Lead Product Designer

Team

CEO, 2 Developers

Impact

Reduced Setup Time for Users by

85%

Eliminated Dropoffs at Configurations by

70%

Decreased Support Tickets by

40%

Research

With limited direct user access and a tight timeline, I started by conducting a comprehensive UX audit of the existing interface and competitive analysis. I collaborated with the founder and developers to understand internal perspectives, but realized I needed to speak with real users to uncover their actual pain points and workflow challenges.

UX Audit. Through a comprehensive analysis of the existing interface, I identified 4 critical issues:

Single-Page Overload

Poor Information Hierarchy

Technical Language Barrier

No Guided Onboarding

Competitive Analysis. I analyzed leading copilot builders (direct competitors) including Microsoft Copilot Studio, Chatbase, and no code tools (indirect competitors) like Claude, and NotebookLM to understand the industry patterns:

Progressive Disclosure

Template-Driven Approach

Immediate Testing

Contextual Guidance

Stakeholder and User Insights. I started by speaking with stakeholders to understand the scope of the problem. I also conducted interviews with 2 program managers from impact organizations.

Developer

"We're spending 4-6 hours daily helping users get through the capabilities configuration. It's becoming unsustainable."

Program Manager

"Users get excited about building a chatbot, but then hit this wall of complexity and either give up or need constant hand-holding."

User 1

"I understand what I want my chatbot to do, but all these technical settings make me feel like I need a computer science degree. I just want to help students get answers about our programs."

User 2

"The interface feels like it's built for developers, not for people like me who just need to get work done. I spent two hours trying to figure out what 'model temperature' means."

Design Principles

Based on the research findings, I established principles to structure the redesign:

Progressive Disclosure

Reveal information gradually to prevent cognitive overload.

Familiarity

Use UI patterns users already know from messaging apps.

Immediate Feedback

Let users test changes in real-time.

Plain Language

Add tooltips and context for technical jargon.

Key Design Decisions

Based on the research findings, I established principles to structure the redesign:

Restructured User Flow with Clear Progression. Transformed the overwhelming single-page configuration into a guided 3-step journey with visual progress indicators:

Addressed the Blank Page Problem. Implemented intelligent prompt assistance through template driven instruction mode, pre-written example descriptions, and contextual guidance.

Persisting Testing Interface. Designed testing as a consistent right-side panel that resembled modern messaging apps.

Other Projects

Gooey.AI: Copilot Builder Redesign

Designed

Gooey.AI: Copilot Builder Redesign

Designed

Gooey.AI: Copilot Builder Redesign

Designed