UX CASE STUDY

Mesh.app

Bridging the Gap Between "Asking AI" and "Getting It Done"

Product Image

Timeline

4 Weeks

Conceptual Sprint

Role

Lead Product Designer

Research to UI

Tools

Figma, Miro

OpenAI (Research)

Platform

SaaS Web App

Desktop First

01 // Summary

The Context

We are in the "Golden Age" of Generative AI. Tools like ChatGPT can write code, draft emails, and create strategies in seconds. But a year ago, I noticed a frustrating pattern.

The Problem: The "Execution Gap"

Despite these tools, creators are burning out. AI can generate the plan, but the transition from 'AI text' to 'real work' remains broken. When a chatbot provides a 20-step list, the manual data entry required to move it into a project management tool becomes so overwhelming that many abandon the project entirely.

The Goal:

Transform the AI experience from a passive "text generator" (Consumption) into an active "visual workspace" (Co-Creation).

Failure Rate
70%

of digital transformation projects fail due to lack of adoption and behavioral change.

Abandonment
42%

of users abandon AI tools after month one due to workflow friction.

"I have 50 great ideas stored in ChatGPT, but I haven't started a single one."

02 // Desk Research & Findings

I conducted deep secondary research, including social listening on Reddit and market analysis of current SaaS and AI tools.

S
Persona

Sarah

The Overwhelmed Visionary. High ambition, low patience. Gets "Analysis Paralysis" with long lists.

25-35 Entrepreneur
x

Transfer Friction

Manual effort copying text from Chatbot to Trello causes major drop-off.

Context Amnesia

Chatbots "forget" project constraints as the conversation scrolls down.

Mountain Effect

"Build Website" feels impossible without a granular breakdown.

Success Metrics

Business Goals

Targets defined for launch.

Retention 20% Increase Completion
Activation < 5m Reduce Setup Time
Conversion High Increase Trust
Day 1 Metric 1st Visual Plan Session 1

03 // Competitive Analysis

The Titans (PM)
Jira, Asana, Monday
High Structure, High Friction. Requires heavy manual setup.
The Generators (AI)
ChatGPT, Claude, Gemini
High Creativity, Low Memory. Text-heavy and disorganized.
The Automation
Zapier, Make
High Logic, No Interface. Connects apps but doesn't guide.

The Blue Ocean Strategy

The market is split.

Miro

allows you to Think (Whiteboard).

Jira

forces you to Manage (Tickets).

ChatGPT

helps you Generate (Text).

There is no "Transition Tool." No tool combines the visual freedom of a whiteboard with the execution logic of a project manager. AI Canvas fills this gap by being a "Generative Workspace."

04 // Ideation

Concept: Auto-List

Failed

AI fills a spreadsheet. Still felt like "Admin Work". High cognitive load for the user.

Solution: A "Living Map"

Selected

Node-based network. Matches mental models. Progressive disclosure hides details until needed.

Core Metaphor NODES not lists

05 // Core Features

Pre-Mortem Interview

Chat interface asks 3 key questions (Budget, Time, Platform) before starting.

Solves: Context Amnesia
Business Goal: Decision Confidence.

Generative Node Map

AI visualizes the project as connected bubbles instead of a vertical list.

Solves: Analysis Paralysis
Business Goal: Activation (< 30s).

Recursive Breakdown

Click a node to "Zoom In" and use AI to split that task into sub-tasks.

Solves: Mountain Effect
Business Goal: Retention.

Agentic Actions

AI stays in the node to perform actions (e.g., "Draft Email," "Search API").

Solves: Transfer Friction
Business Goal: Engagement.

06 // Interaction Flows