Product Exploration
Overview
Product Designer
4 weeks
Hosted on Maven
Become an AI Product Designer
Main Challenge
Many DIYers feel motivated by ideas they see online, but struggle to get started. Tutorials are scattered. Skill levels are assumed. And decisions — from tools to steps to materials — often paralyze momentum.
In this project, I set out to understand where people get stuck, and how AI could help organize and elevate their creative process without taking over.
Research
I conducted a deep dive into the current landscape of AI product design. The goal was to understand how AI is being integrated into digital tools today—both successfully and unsuccessfully. I analyzed a range of AI products from chatbots like Perplexity to tools like Notion AI, Uizard, Lovable, DOT and Midjourney—to understand how AI is being positioned within the product experience. Through this research, I identified key friction points and ethical considerations designers face when working with AI.
Painpoints
Users often don’t know where to begin once they have an idea.
“I have the idea... but I never know where to start.”
Materials on hand are rarely factored into project discovery.
“I’ve got all this scrap wood—what can I even make with it?”
There’s a desire for AI to act as a collaborator, not a replacement.
“I don’t want AI to take over—I want it to help me build.”
Design Methods
Smart Recombine is a creative method for reimagining interactions by merging them with ideas from other disciplines. I was introduced to this approach by my instructor, Maheen Sohail, during the "Become an AI Product Designer" course. For this project, I applied the method by blending the exploratory nature of moodboarding with the structured logic of visual scripting.
Moodboarding
Visual Scripting
Solution
Buildable is an AI-powered DIY workspace that turns scattered ideas into structured projects. Instead of treating AI as a passive assistant, I reimagined it as a collaborative workspace partner. One that helps users narrow down ideas, plan projects, adjust based on skill level and materials, and feel supported from start to finish — all without assuming too much control or adding complexity.
Consideration #1
Users can start by chatting with the AI or uploading photos of materials they already have. The AI then asks guided, contextual questions to help narrow down ideas and generate project suggestions. Conversations are saved and can be referenced at any time, supporting continuity and long-term creativity.
Consideration #2
If AI-generated suggestions aren’t quite right, users can draw a rough sketch with notes. Buildable’s AI interprets these to create visual concepts that are more aligned with the user’s vision. It bridges the gap between low-fidelity ideas and high-fidelity previews.
Consideration #3
Once an idea is formed, Buildable auto-generates a comprehensive overview. This helps users assess feasibility and prep everything upfront.
Consideration #4
Buildable adapts in real-time. As users interact with the overview (e.g., selecting “Materials”), the AI branches out into relevant subcategories—such as a curated shopping list based on availability and price. Each decision unlocks deeper customization.
Consideration #5
A persistent chatbot sits alongside the workspace, offering real-time support. For example, if a user says, “I want to add a navy stripe to my desk,” the AI suggests matching paint tones and updates the project preview accordingly—merging design feedback with visual iteration.
Consideration #6
At the end of the journey, Buildable packages everything—plans, sketches, instructions, materials—into a neat, shareable file. Users can zoom out and reflect on their entire creative process, or share it with friends, family, or the Buildable community. A future iteration could allow Buildable to format these for remixing, helping others adapt and build on finished projects
Outcome