Re-design the AI campaign builder to achieve self-service and better user adoption (In-progress)
The AI agent that handles email outreach for creators and SMBs while they sleep
Creators and small businesses rely on email to engage with their audiences. But personally reaching out to prospects is time-consuming and unsustainable.
With an autonomous AI agent that handles email outreach for them, our first batch of customers received significant gain in their revenue.
Now the team is bringing the product experience to the next level.
Problem: Customers won't create new campaigns themselves.
We found many customers loved the results but still leaned on our team to set up their outreach campaigns.
As the founder pushing more customers to use the self-serve builder tool, we've heard from many that the existing flow was not easy to get hands on.
"The sales agent was great, but I just don't have the time and energy to learn that complicated set up."
Core issue: The System-oriented UX is blocking user adoption.
The original workflow creation flow was designed around what the AI system needed, not what users understood.
Customers were front-loaded with questions and configuration fields before seeing any output.
How might we make the campaign creation UX speak to users, not systems?
My solution: Same inputs, reframed with grounded, prioritized, and user-centric UX
1. Set up a clear mental model with hierarchy, progress, and user-friendly copy.
2. Make AI reasoning transparent and explainable, framed with user value not data schema.
3. Remove blank canvas frictions with prompt assists. Prioritize required fields and save optional for later.
Establish a Minimum-Viable Design System with Claude Code to iterate fast and cohesively
Brand Typeface
Tech Gadget
Carbon Design System
We use this first redesign project to build a lightweight design system as the starting point.
I drew inspiration from Constant Closer's brand, IBM's Carbon DS, and the aesthetic of tools like Teenage Engineering: minimalist, modular, clean.
The system evolved alongside the design. Components were added and refined as the work progressed.
Building Now
Campaign workspace: a central hub to view, refine, and manage
The next phase is building the campaign workspace for users to review their generated email sequences, edit follow-up timing, and adjust agent behavior or advanced settings without engineering support.
🫡 To be continued...
Takeaway
Understanding the technical framework is the design work, not a prerequisite to it.
I invested a lot of time in mapping the AI system's logic and from there I created the right mental model, aimed to let users build the campaign effectively without the cognitive barrier.
The design system and Claude Code prototyping allowed me to move fast, stay consistent, and inspect the details that static Figma files couldn't catch.