Keyword Insights
Took an AI SEO tool from spreadsheets to a product: one design system for about ten features, and a new onboarding.

- Role
- Product designer, founding team (part-time)
- Timeline
- May 2022 – Aug 2025
- Scope
- SaaS · AI
- Turned a spreadsheet-driven tool into a product with one design system
- Grew the system to hundreds of components across ~10 very different features
- Onboarding v3 joined separate features into one flow and stayed the main onboarding
Keyword Insights helps content teams decide what to write next: it groups search keywords into topics and plans the content around them. When I joined in 2022, the product was spreadsheets and exports with no interface at all. The team was two founders with an SEO background, two back-end developers and one front-end developer, and nobody wrote specs for design. So I defined the interface myself and turned the spreadsheets into dashboards.
From spreadsheets to a product.
Design system first
With about ten features waiting for an interface, designing them one by one would have made them look like ten different products. I built the design system first and pulled each feature onto it. The features had little in common, from keyword research to clustering and content planning, so the library grew to hundreds of components, and keeping it consistent enough for developers to build without asking me was most of the job.
Animated feature illustrations I made for the product, one per feature, in the system’s palette.
Part of the component library.
Six features built on the same system.
Onboarding
The product grew into several SEO features that belonged to one workflow but felt like separate tools, and new users didn’t know where to start. Two onboarding experiments didn’t fix it. The third did. It joined the features into one path: a step-by-step tour showing what each feature was for, and highlights with short instructions on the components people found confusing. Activation and retention went up, though I have no numbers to show, and v3 stayed the main onboarding.
The version that never shipped
The next step was an internal product that hid every feature behind one flow: from a business’s name to ready articles in a couple of clicks, at a fixed price. It was designed, then shelved when the founders disagreed on direction.
From a URL to an order of ready articles.
Beyond the app
The same role included client work: I designed and redesigned client websites and advised clients on improving the ones they had.
Client websites.
Best designer and product manager out there. Been together through numerous projects from MVP to scaling to profitability, Anton does it all. Top responsibility and communication skills.
— Mykola M., CTO, Keyword Insights
The team
It was hard without a PM. The team spoke different languages, and we never had a shared vocabulary for the product. Everyone was friendly, but there was no core holding it together. We still keep in touch, and most of those people now work with cashtime in one way or another.
What I’d do differently
- Plan for design system v2 earlier. As the only designer I built for the product we had. By the end it needed a second version I never got to.
- Don’t go part-time on a team that needs full-time speed. Part-time, I couldn’t deliver as fast as the founders wanted, which is why I left in 2025.