Scatterwise

Inspiration

A lot of productivity and brainstorming tools assume that people think in a linear way: come up with an idea, organize it, make a plan, and then start working.

That isn't how I think, and it isn't how many of my neurodivergent friends think either. When I'm working on a large project, I often have a lot of ideas and connections, but they don't arrive in a useful order. The difficult part isn't necessarily coming up with ideas — it's turning all of those disconnected thoughts into something I can actually use.

That inspired me to build ScatterWise.

Instead of asking users to organize their thoughts before they can be productive, ScatterWise starts with the mess.

What it does

ScatterWise lets users quickly dump their thoughts as individual notes without worrying about structure, order, or how they connect.

AI then analyzes those thoughts, identifies relationships between them, and decides how to organize the information into a useful visual structure. Depending on what the user is trying to accomplish, that could mean turning scattered ideas into a connected word web, an organized hierarchy, a process, or another visual representation.

The goal is to bridge the gap between having ideas and being able to act on them.

Rather than forcing neurodivergent users to adapt to a traditional productivity system, ScatterWise adapts the system to the way the user is already thinking.

How I built it

I built ScatterWise as a web application with an interface designed around making idea capture as frictionless as possible.

The workflow is intentionally simple:

1. The user enters thoughts as they come to mind.
2. ScatterWise sends those ideas to an AI model for analysis.
3. The AI looks for relationships, categories, dependencies, and patterns between the ideas.
4. It determines an appropriate way to structure the information.
5. ScatterWise turns that structure into an interactive visual result that is much easier to understand and use.

A major part of development was figuring out how to use AI for more than simply summarizing text. I wanted the model to actually reason about the relationships between ideas and help determine how the information should be represented.

Challenges

One of the biggest challenges was dealing with how unpredictable brainstorming can be. Users might enter complete sentences, single words, tasks, questions, or ideas that only make sense when connected to something else.

That meant I couldn't rely on a fixed template or assume every brainstorming session should become the same type of diagram. I had to design ScatterWise so the AI could interpret the information first and then determine how it should be organized.

Another challenge was keeping the interface simple. Adding more menus, categories, and configuration options could technically give users more control, but it would also recreate the exact problem ScatterWise is trying to solve: requiring users to organize everything before they can get started.

What I learned

Building ScatterWise changed how I thought about using AI in accessibility tools.

I learned that AI can be especially useful when it acts as an adapter between the user and traditional software. Instead of expecting every person to change how they think to fit the structure of an application, AI can interpret how someone naturally communicates and transform that information into a form traditional tools can work with.

I also learned that designing for neurodivergent users isn't just about adding accessibility features after building a product. It can mean reconsidering the basic assumptions behind the product itself.

ScatterWise started with one question:

What if you didn't have to organize your thoughts before your tools could understand them?

Built With

• ai
• codex
• next.js
• node.js
• tailwind
• typescript
• vercel