Close

Proma

Proma's founder found me on LinkedIn and asked me to elevate its visual storytelling. That one ask turned into a new brand, a new website, event material, the product's first design system, and the first time AI became a real part of how I design.

  • Brand strategy
  • Brand identity
  • Web design
  • Event collateral
  • Design system
  • Product design
  • AI design workflow

I had just redone my portfolio and posted it on LinkedIn when Shishir messaged me. He had a product, and he wanted to "elevate its visual storytelling." Those were his exact words.

His company, Growthzilla, already had a product doing well in the US, called Uzeli. Proma was the next one, and it needed a story.

Before a story, though, it needed a brand, because there wasn't one. The logo was one thing, the product looked like something else, and the engineers were each doing their own thing. No design system anywhere. Everything was in shambles, which is exactly why he reached out.

So we went straight to branding. We would get on long Slack huddles and he would tell me how he saw Proma, where it came from, where he wanted it to go. I listened and built the picture in my head. Then came brand nouns, a lot of logo sketches, and a lot of back and forth.

Hectic is the right word for that back and forth. He's a perfectionist. I'm a perfectionist. Two perfectionists looking for the perfect logo takes a while.

But it was doing something useful. When we started, neither of us actually knew what story Proma should tell.

We figured out what Proma needed to be somewhere in the middle of discussing its logo.

"Proma will be a composable enterprise platform." That's where Shishir landed. He also always talked about four pillars that held up the product's philosophy. So the logo became those four pillars, as composable modules. That's when the logo clicked.

How the Proma mark is built: the Roman pillar, composability and interconnected nodes come together as one icon

We were excited, and we had no time to enjoy it, because an event for Proma was coming up fast. The website was in the same state the brand had been in, so I revamped it around the new story. I also made the physical things for the event, the flex banners and posters.

Proma event posters
The Proma website on a laptopPages of the Proma website

With the brand and the website sorted, the product was the last thing still in shambles. No designer had ever touched it, so there was a lot to fix. I built a design system in Figma that every engineer could actually use, then reworked flows and a lot of screens. The front-end engineer had a design background, which made it easy. I would bring ideas, we would talk them through in the same language, and iterate fast.

  • Proma product screens: home, creating a view, invitations and the form designer, 1 of 4
  • Proma product screens: home, creating a view, invitations and the form designer, 2 of 4
  • Proma product screens: home, creating a view, invitations and the form designer, 3 of 4
  • Proma product screens: home, creating a view, invitations and the form designer, 4 of 4

Right around then, Claude showed up, and iterating got even faster.

I moved the design system from Figma into Claude. Then I wrote MD files with everything that goes into my process: the philosophy, the steps, the kind of references I look for. After that, prototyping got very fast. I prototyped, and the engineers built it.

Soon the engineers were using Claude too, and that changed the handoff. For screens that needed a revamp, I started writing one MD file of guidelines. How each screen should behave, which interaction for which case, what to avoid. I attached a few Figma screens for reference, and the engineer just gave it all to Claude.

Design is basically a list of what needs to be done and what doesn't. So I wrote the list down.

The whole settings section was rebuilt from one of those files. That was the first time AI was part of how I design, not just something I had tried.

NextDMC Lounge