From conversations to command center

Improving Chat Management on ChatGPT through Workspace-Style Layout and Smart Navigation

Methods

Desk Research

User Interviews

Contextual Inquiry

User Persona

Competitor Analysis

Impact–Effort Matrix

5 Second Test

Our Team

2 Product Designers

My Role

User researcher, Visual designer

Duration

May 2025-July 2025

Our Mentor

Marzie Nadali

Introduction

We’ve all been there.


You remember a brilliant answer from GPT…

But now you’re endlessly scrolling, trying to find it.
That’s exactly why we set out to fix it.

We were frustrated too!
finding old chats felt harder than it should. That’s exactly why we decided to tackle this problem.

What are we trying to understand?

The Smartest Assistant Still Needs Smarter Organization !

We conducted field research across platforms like Twitter, Reddit, and YouTube comments to uncover users’ biggest frustrations with chat management in ChatGPT.

Key Repeating Patterns

What patterns keep showing up across user feedback?

Even with the launch of Projects, users still face challenges. Whether on Reddit or Twitter, people consistently highlight gaps in how ChatGPT supports continuity, organization, and memory.

Watching Real People Work

Validating Patterns Through Real Conversations

We ran interviews and contextual inquiries to see if the issues we found in online comments showed up in real use. Watching how people use ChatGPT in their daily flow helped us spot what truly gets in their way.

Assumptions Meet Reality

What We Got Right And What We Didn’t

Our interviews and contextual observations helped us separate facts from guesses. What do users actually need, and what did we just assume.

Who Are We Designing For?

From Casual Prompts to Project Workflows: GPT Usage Patterns

We identified two key user types, casual and professional. Each with distinct needs and friction points. By mapping their usage habits and frustrations, we uncovered patterns that shaped our design direction.

Competitive Analysis

What Others Do Well And Where AI Chats Still Fall Short

We compared ChatGPT with other AI Chatbots and productivity tools to see what features they’ve nailed and where they still struggle. This helped us identify inspiration points as well as clear opportunity gaps.

⚠️ Where Everyone Loses Track

Most AI chat tools like ChatGPT, Claude, and Copilot are fast but unstructured:

  • No folders or grouping to structure projects

  • Poor in-chat search, so old insights get lost

  • Limited chat management, making long-term work a mess

✨ Where Others Shine

Apps like Gemini, Poe, and Coda Ai are already tackling organization by offering:

  • Privacy-first sharing (customizable link permissions)

  • Pinned chats (keep key threads front and center)

  • Clean timelines (simple browsable history even in free versions)

Design Decision

Turning Frustrations into Design Opportunities

We compared ChatGPT with other AI and productivity tools to see what features they’ve nailed and where they still struggle. This helped us identify inspiration points as well as clear opportunity gaps.

Prioritizations Matrix

Which solutions matter most to each user?

Although John and Sara face many of the same struggles, they use ChatGPT in different ways. We prioritized ideas based on their unique workflows, so each one gets the most out of their experience.

Turning Ideas into UI

Designed for How People Actually Work

We wanted to fix the little things that were quietly frustrating people every day, like lost files, forgotten chats, and not knowing where to click next. Every small change we made was based on real problems users were facing.

✨ A Homepage That Works Like a Workspace
✨ A Homepage That Works Like a Workspace
✨ A Homepage That Works Like a Workspace

  1. Dashboard View – Card Layout

  1. Dashboard View – Table Layout

  1. Workspace View – Expanded Sidebar

✨ From static folders to dynamic project tracking
✨ From static folders to dynamic project tracking

  1. Project Overview

  1. Project Timeline

  1. Project Files

✨ Search smarter, navigate faster
✨ Search smarter, navigate faster

  1. Smart Tags for Faster Search

  1. See Results with Highlighted Keywords

  1. Jump to Results Inside the Chat

✨ Highlight what matters, save it where it belongs.
✨ Highlight what matters, save it where it belongs.
✨ Highlight what matters, save it where it belongs.

  1. Jump to Results Inside the Chat

  1. Move Highlights to Collections

Validating with users

Where We Nailed It & Where We Missed

We showed our design to 5 previous interview participants for 5 seconds and asked quick, targeted questions to understand if core features and purposes are immediately clear. Their reactions helped us spot what felt intuitive and where more clarity or refinement was still needed.

We showed our design to 5 previous interview participants for 5 seconds and asked quick, targeted questions to understand if core features and purposes are immediately clear. Their reactions helped us spot what felt intuitive and where more clarity or refinement was still needed.

Success Metrics

Did our features actually help people work smarter?

We had a few key metrics in mind that we would’ve tracked if we had the opportunity. These reflect the core values of our design such as clarity, usability, and long-term user value.

We had a few key metrics in mind that we would’ve tracked if we had the opportunity. These reflect the core values of our design such as clarity, usability, and long-term user value.

As we conclude...

Even the smartest tools, like GPT need thoughtful design to truly serve their users. While these AI systems are fast and capable, we learned that intelligence alone isn’t enough. When users spend hours daily in a tool, it must evolve into a real workspace, not just a chatbot answering questions. Our goal wasn’t to add more tiny features. It was to rethink the workspace entirely. We focused on turning GPT into a place where people can work, organize, and stay focused ,without jumping between tools. Because when AI feels like a true workspace, it’s not just powerful. It’s personal, practical, and productive.

Even the smartest tools, like GPT need thoughtful design to truly serve their users. While these AI systems are fast and capable, we learned that intelligence alone isn’t enough. When users spend hours daily in a tool, it must evolve into a real workspace, not just a chatbot answering questions. Our goal wasn’t to add more tiny features. It was to rethink the workspace entirely. We focused on turning GPT into a place where people can work, organize, and stay focused ,without jumping between tools. Because when AI feels like a true workspace, it’s not just powerful. It’s personal, practical, and productive.

Crafted with too many Figma tabs and a dangerously low battery.

©2024 Portfolio by Elly khajooee

Other Works

Resume

Email

Linkedin