For Kyle at Conversion
A design system built for Conversion's complexity
I spent some time studying your product and I kept thinking the same thing: this is a really ambitious platform with a lot of moving pieces. Here's what I think a design system built specifically for Conversion could look like.
What I see
There's a lot going on here (in the best way)
I dug into conversion.ai, read through the TechCrunch coverage of your Series A, and went through your LinkedIn posts. A few things kept jumping out at me.
6+ product surfaces,
each with its own world
Email Studio, Automations, Data Model, Lead Scoring, Forms and Pages, Reporting. Each one has different interaction patterns, different data density, different user expectations. Keeping all of that consistent across a fast-moving product is a real challenge.
You're replacing tools
people have used for years
Your customers are coming from Marketo, HubSpot, and Pardot. They're expecting enterprise polish and patterns they recognize, while you're building something completely new. That's a tough balance to strike without a system behind it.
AI agents are a totally
new pattern space
Campaign Agent, Analysis Agent, Personalization Agent, Audit Agent. None of these exist in legacy MAPs, so there's no playbook to reference. Each one needs patterns for status, confidence, human oversight, and results. That's a lot of new UI to figure out as you go.
You're growing
really fast
Nearly 8 figures in revenue, 4,000+ customers, displacing some of the biggest names in martech. At that pace, every new feature ships faster when it can pull from a shared system instead of being designed from scratch each time.
What I'd build
A system designed around what Conversion actually needs
Not a generic component library. I'm talking about a system built around the specific patterns your product uses every day, designed to grow with your team and your roadmap.
01
Data-dense component foundations
Conversion lives in tables, records, and dashboards. I'd build a component system specifically for your data patterns: unified record cards, score displays, status indicators, sync state badges, metric tiles. These are the building blocks that show up on every single screen of your product.
02
Workflow canvas system
Your visual workflow builder is the heart of the product. I'd create a pattern library for nodes, branches, conditions, delays, and AI actions that scales as you add new workflow capabilities. Consistent enough to feel unified, flexible enough to handle whatever automation logic you throw at it.
03
AI agent interaction patterns
As you ship more agents, a reusable agent UI kit would give you consistent patterns for showing work in progress, displaying confidence levels, surfacing insights, and reporting results. So every new agent feels immediately familiar to users without reinventing the wheel each time.
04
Token architecture
Colors, spacing, typography, elevation. I'd set up a token structure in Figma that maps directly to your codebase so designers and engineers are always on the same page. No more eyeballing spacing or guessing which blue to use.
05
Contribution model and documentation
Not just documenting what exists, but capturing why decisions were made. Pattern guidelines, do's and don'ts, and a contribution model so anyone on the team can extend the system without breaking it. The kind of foundation that makes every new person productive faster.
Why now
Post-Series A is when this starts to really matter
Before Series A, speed matters more than consistency. You're finding product-market fit, shipping fast, making decisions on the fly. That's exactly what you did and it clearly worked.
After Series A, things shift. You have 4,000+ customers expecting enterprise polish. You're going head to head with Marketo and HubSpot on UI quality. You're shipping AI agents that need entirely new interaction patterns. And every new feature is touching components that were built quickly during an earlier phase.
This is the point where a design system stops being a nice-to-have and starts being the thing that determines how fast your team can ship without accumulating design debt. I've seen both sides of this play out, and the companies that invest here now are the ones that pull away from competitors over the next year.
Recent work
This is what I do
I build design system infrastructure for scaling B2B products. I work directly with product leaders and designers who need their teams to ship faster without things getting inconsistent. Here's what that looks like.
Whisker Labs (Ting)
Currently building design system infrastructure for their consumer product, working directly with their Director of Product. Complex platform with data visualization dashboards, alerting workflows, real-time grid monitoring, and device management across mobile and web.
Thalamus
Built design system infrastructure for a data-heavy B2B platform, working alongside their product designer. Complex institutional workflows, multi-role dashboards, and a growing product surface that needed consistent patterns across features.
SMB.co
Built their first design system from scratch, giving the team a foundation they could actually build on as the product grew.
"Keara built our first real design system and gave us a foundation we could truly build on. It was exactly what we needed at that stage."
Next step
I'd love to walk you through this
No pitch deck, no pressure. Just a conversation about what Conversion's design system could look like and whether it makes sense to work together.
Or just message me on LinkedIn. Happy to chat there too.
hello@keara.design
Email Copied