YaloCode: Making everyone a builder
February 2026 – July 2026 · Senior Product Designer
Overview
YaloCode started as an internal hackathon experiment with an ambitious goal: make Yalo's platform accessible to the teams responsible for building and delivering conversational agents, without requiring deep technical expertise.
I joined after the initial prototype and worked across the product system, from E2E workflows and information architecture to the behavior that shaped how YaloCode responded. I defined reusable skills, behavioral instructions, response structures, and the rules that determined when the product should respond conversationally or shift into a more structured interaction. I also designed the chat patterns and components that supported those behaviors, and contributed directly to frontend implementation.
YaloCode became a shared workspace across Yalo, supporting Customer Success, Sales, Conversational Design, Engineering, and external implementation partners. Today, it is actively used by around 80% of the company, with a 4.5/5 CSAT, while WhatsApp agent delivery has gone from as much as six months to under two weeks.
Marketing footage courtesy of Yalo
Chat UI Gallery
As YaloCode moved into real workflows, we noticed a trust gap: users would ask the agent to do something, but sometimes complete the task themselves because they weren't fully confident in what it would do.
We introduced structured widgets that made planned actions visible and reviewable, giving users a chance to understand what would happen next and decide whether to proceed.
Below is a recreated preview of selected interface patterns.
Agent creation
Want me to guide you step by step?
I'll ask you a few quick questions to set up the ideal sales agent for you
Designing for builders
Once users could see and review what the agent was doing, the next opportunity was making those workflows easier to reuse, share, and build on.
People were already creating their own ways of working across tools like ChatGPT, Claude, Gemini, and Yalo's platform. Instead of asking everyone to start from scratch, we turned those behaviors into reusable building blocks. Skills could be created and shared, eventually growing into a marketplace, while the chat gallery made useful workflows easier for others to discover.
What I liked most about this direction was that people didn't have to figure everything out on their own. They could start from something that already worked, adapt it to what they needed, and eventually create something others could use too.
