Userbot: AI Agents
Leading the Shift to AI Agents
As a freelance Product Designer, I supported Userbot’s transition from manually trained, rule-based chatbots to an LLM-powered AI agent platform. I redesigned key experience flows and introduced AI-native features that enabled users to configure goals and behaviors instead of scripting conversations, making advanced automation more scalable and intuitive.

Overview
Userbot embarked on a strategic redesign to integrate LLM capabilities at the core of its product, redefining how conversational automation was built and managed. The project involved reshaping the platform’s architecture and user experience to support more intelligent, flexible interactions while maintaining control, reliability, and clarity for business teams. My role focused on translating this technological leap into a coherent, scalable product experience.
Unlocking AI Agent potential
Knowledge Base
I introduced a centralized Knowledge Base where users can upload and manage documents that power agent reasoning and contextual understanding. The main challenge was designing multiple document-specific setup configurations while maintaining a consistent, scalable interface that keeps complexity under control and ensures long-term flexibility.

Agent & Tools
I redesigned the Flow Builder to support modular agent blocks such as routing, retrieval, Q&A, and variable extraction, and introduced a scalable “toolbox” layer of agent tools to extend specialized capabilities within workflows. The core challenge was translating abstract agent logic into clear, consistent interaction patterns, ensuring the system could handle complexity while remaining intuitive and extensible over time.


Reimagining Core experiences
I reimagined the Flow Builder’s core structure and commands, repositioning and redesigning key actions such as navigation, preview, save search, and workflow management. The challenge was making these complex interactions intuitive and consistent, ensuring that both existing manual bots and new AI agents could coexist seamlessly while keeping the system scalable and user-friendly.

I redesigned the Workspace page, where users select and manage their bot or AI agent projects. The challenge was presenting both legacy bots and new AI agents clearly, making it easy to switch, access, and manage projects in a scalable and intuitive way.
