ArtCenter Entrepreneurship Course

Ideator

UX/Interaction Design
AI Ideation Workflow
Design Systems
Visual Design

Timeline: Jan 2023 – Jan 2024
Team: Solo project
Role: Product Designer & Founder

An AI ideation toolkit that helps students get unstuck by pairing proven UX methods with image-to-text AI.

Ideator brings popular UX methods like SCAMPER and Mashup into one workspace. Students can start from a method or a blank space, turn an inspiration image into keywords with AI, and chain methods together to push an idea further.

I built Ideator on my own in an ArtCenter entrepreneurship course, where the goal was to launch a real product. Interviews with 15 designers, students, and business professionals pointed to three blockers: ideating alone, not knowing where to begin, and not knowing where to look. I turned those findings into wireframe prototypes, tested them with 20 people, compared two image-to-text AI models, designed the final product and its style guide, and pitched it at the end of the course.

If you’d like to learn more, get in touch.

HERO ARROW ACCESSIBILITY FIX (any case study) Paste this at the end of the case study's intro code block. Dreamforce already includes it. Change the outline color to that case study's accent text color if you like.

The challenge

How might we help students explore new directions when they’re stuck on their next big idea?

Problem

Students struggle every day to find new ideas and solutions for their coursework, and often don’t know where to begin.

Solution

An AI-powered ideation toolbox that sparks inspiration through proven UX methods and design thinking.

Research

Why students get stuck

Students are under real pressure, and new ideas don’t come easily when you’re overwhelmed.

21%of students felt unprepared for college courses, up from 11% the year before
58%felt stressed or overwhelmed by time spent studying and supplementing their learning
1 in 3considered dropping out of college because of these challenges

Source: McGraw Hill

I interviewed 15 designers, students, and business professionals about how they come up with ideas. Three patterns kept coming up.

Insight 01
They ideate alone first

People like to ideate on their own, then bring their findings and finished thoughts to the group.
Insight 02
They don’t know where to begin

Whether it’s writer’s block or too few perspectives, getting started is the hardest part.
Insight 03
They don’t know where to look

With an overwhelming amount of information available, it’s hard to know where to find what’s useful.

The solution

One workspace for guided ideation

Ideator pairs popular UX methods with AI so students always have a next step. Here are two short walkthroughs of the final prototype.

Start from a method Open a space and work through each SCAMPER step.

From an image to new ideas Generate keywords from a photo, then chain a method on top.

01
Start from a method, not a blank page

Start with an empty space, or quick-start a session with a popular UX method like SCAMPER or Mashup.
02
Explore different directions

Each tool comes with a set of methods to apply to your thinking. SCAMPER alone has seven prompts, so one idea can branch in several directions.
03
Turn an image into ideas

Drop in an image that inspires you, and Ideator pulls out keywords and themes to build on.

1Add an image and press Generate.

2Ideator returns keywords and themes to build on.
04
Chain methods together

Layer UX methods on top of earlier results, like running SCAMPER on keywords from an image, to reach more unexpected ideas.

1Pick a method, like SCAMPER, to run on those results.

2Substitute suggests new materials, ready to chain again.

Process 01

Choosing an image-to-text model

Image to text had to turn a single photo into keywords worth building on. I ran the same image through two models to compare the results.

CLIP Interrogator

Hugging Face

Pro
Wide range of results, grouped into categories

Con
Queries can take a while
Replicate

Replicate API

Pro
Fast output

Con
Less detail and a narrower range of results

The comparison came down to a tradeoff: richer, categorized keywords to build on, or faster results.

Process 02

What users asked for

I tested wireframe prototypes of the MVP with 20 people, then revised them based on what they told me.

Two MVP wireframes with a 5:00 timer: a blank canvas, and the same canvas with a searchable panel of UX methods

MVP wireframes used in testing

“I want to be able to sketch, write, input various media and potentially links.”
“What if I want to continue to generate based on these?”
“I work visually, what if your system could ideate this way?”
“Can your system support students through suggestions and create a more collaborative experience?”

Two changes came out of testing: chaining methods, so people can keep generating from earlier results, and SCAMPER quick-starts, so people who don’t know where to begin get a first step.

Visual design

A style guide for the whole toolkit

New Spirit sets the display and headline styles, and Satoshi handles titles, labels, and body text. Ten “Inspire” accent colors, from InspirePink to InspireBrown, give each method and step its own color, alongside neutral fills and semantic colors for alerts.

Ideator style guide: New Spirit and Satoshi type scale, ten Inspire accent colors, neutral fills, text colors, and semantic colors

Positioning

Where Ideator fits

I mapped existing ideation tools from prescriptive to imaginative and from fixed to modular. IDEO Method Cards, IdeaFlip, and most whiteboard tools lean prescriptive.

Ideator aims for the modular, imaginative corner: for students, professionals, and thinkers who want an imaginative, intelligent, and collaborative way to ideate.