Tools I Can’t Design Without

A friend of mine recently asked me what tools I’m using for product design. I’ve talked about this somewhat in bits and pieces, but it’s been a little while since I’ve captured it all in one place. Plus there’s been some changes recently.

Following are what I use to take design ideas from initial inception all the way through to shipping. Some of this is very much related to my professional work, but other parts are valuable just for how I think about processing information.

Freeform

This is still the best tool to sketch out ideas and figure out how they all fit together. If you need to process information in anything outside of text, and you want a way to let it all free float and connect ideas with an Apple Pencil, then this is the best app to use. I’ve tried Miro, Linea Sketch, Figjam, and numerous other white boarding apps, but each has some flaws and limitations that make the thought process of diagramming and sketching cumbersome.

Of course there’s pencil and paper, which I’m quite fond of at times for their ability to remove the interface limitations and allow for free-flow thinking.

But I’ve grown fond of cut and paste, changing colors, moving elements around easily, and saving all the boards into a single device that I can carry with me.

Freeform has its limitations, and the latest iPadOS betas almost made me switch to other apps, but far and away its the best tool for this purpose.¹

Figma

I’ve had a love and hate relationship with Figma over the years. It’s the best tool available for taking Freeform sketches and bringing them to mid-fidelity or high-fidelity prototypes. The Figma Mirror app on iPhone is underrated, and one of my favorite tools for getting a true “feel” for how a mobile app functions in hand. Creating clickable prototypes feels like a superpower, where you can tap around on different parts of the screen and have it come to life in realtime.

It doesn't, of course, any sort of programmatic process to create things for you (although I'm curious if Figma Make is getting any good), so every screen has to be meticulously laid out. But sometimes a higher level of fidelity is necessary to know if a design is going in the right direction. And when showing concepts to users or stakeholders, you can gauge if it makes sense as they’re attempting to navigate through your screens.

ChatGPT

For nearly every step of the design process I'm now tossing ideas and questions into ChatGPT. If I’m stuck on a component, I’ll typically go off to Mobbin or Google to scour through and find inspiration, but now I'm testing asking the chat bot for feedback, to help identify what’s working and not working with the idea.

With an interview loop, I can often pull useful information from ChatGPT, far better than simply asking a question or two. This process now happens daily, and often only takes 5-10 minutes each time, quick enough to pause and actually write out what I’m trying to accomplish.

I’m still finding that it fails at generating anything visual with UX or UI design. Its attempts are completely wrong most of the time, and its code results aren't reliable either.

But for defining a problem, presenting a written solution, and helping identify the hierarchy of how things fit together, it's incredibly valuable.

Grok

Last month I tested Grok alongside ChatGPT with nearly every prompt, to see how much it offered as an alternative to the juggernaut. The short answer is it’s good, and often helps uncover pieces that I missed when working with ChatGPT; but because of the extra overhead and hassle I often pause when debating whether to interact with it for a second opinion. Sometimes, when I’m waiting on ChatGPT for an answer, I’ll throw the same prompt into Grok just to see what it offers as an alternative.

But longterm I see myself canceling the subscription, at least till its value is more apparent.

The one area it’s better at is more in-depth work tasks. If I ask for a table of data based on a specific request, and I need hundreds of results back, it’s likely going to keep trying longer and will give me something closer to what I’m wanting.

Claude Code

With the launch of ChatGPT 5 I’ve started re-testing the flow of sketch > code. So this might change in a bit next time I give an update.²

But for now, Claude Code is my favorite for taking rough Freeform ideas and running them to a coded prototype. It takes a little longer to setup, but once it’s running as a companion app to your local browser installation, you start to feel like you have a super-powered buddy helping you along.

For someone who is very nervous to use the Terminal, it’s made that whole process much easier (and potentially more risky).

My ultimate plan is to use this, or something similar, to replace Figma. That probably won't happen in 2025. I can’t do it yet, there’s still too many useful pieces of Figma I rely on to quickly ideate. But longterm I hope to go straight from sketches to code, via an LLM of some sort.

Notes

I’ve written before about my note capturing process, the only thing I’ll add is the addition of Tot. While listening to one of the Relay podcasts, I was inspired to try it again for the fourth time. And it's stuck.

Though I rely heavily on Apple Notes, I often spin up TextEdit files to create temporary, desktop saved, text files. These are for quick meeting notes, or todos, and at the end of each day I sort them into Things 3 or Apple Notes (or Notion, or Linear, or Figma). These temporary files are my version of sticky notes, and quite useful.

Also, because I keep an empty Mac desktop (apart from when I’m in an active project for the day), the notes stick out and remind to come back and process them as outstanding items.

At this point I’ve been testing Tot over the last few weeks to see where it can fit in the tools I use. So far I’m finding it useful for notes that are slightly more permanent than TextEdit, but where I also don’t want them lost in the hundreds of Apple Notes (I’m bad at using anything other than search, scrolling to the top for recent, or pinning anything). My newsletter articles start in Tot now (although I’m finding some issues with the Tot > Ghost Markdown copy+paste flow), and some not-urgent ideas sit one of the eight dots as well (such as things I want to update on the Trailblazers Trivia).

Putting it all together

When I sat down to write this I thought it’d be more complex, that I’d describe a diagram of dozens of apps all fitting together like Tetris pieces. There’s more I use, of course. Lots of Safari and Chrome tabs float around on a daily basis, I use Slack, check email, Calendar, and a hundred other things.

But when it comes to the process of trying to design an app, these tools are enough. Much of the actual work happens in different ways, through collaboration, feedback, thinking, real world testing, and understanding the principles of good design through the school of experience. The tools often fade into the background to make way for thinking. And, at least for now, that's a good thing.

1. Of course you can’t rotate objects in Freeform (or if you can, I haven’t figured out how), resizing is a nightmare (it works sometimes, and other times all the disparate pieces will shuffle across your board in unpredictable ways), and “send to back” and “send to front” are a joke. They work when the software feels up to it, and never when actually needed. But all that crazy, very basic functionality weirdness aside, the tool lets me just sketch and draw and mostly not think about anything else. Also, since switching to the M3 iPad the memory constraint issues I was dealing with (laggy, slow, boards and artwork) have gone away.
2. Yes, at some point all the new LLM model changes are becoming a bit ridiculous. We’ve arrived at a time where major shifts in our technology are occurring on a weekly basis. It’s hard to keep up, but I love this stuff. If you’re following along and thinking, “oh I thought XYZ model was the best”, it’s probably just fine for the type of work most of us do. And in reality I don’t push any of these LLMs even remotely close to what they’re capable of, most of the time. But in testing this product design flow, I’m constantly checking just to see if something has changed, poking around to see what works best. I might even try UX Pilot again. Because of how fast things are changing, you may see different recommendations from me on a monthly basis. I’ll try not to be too crazy about it.