AI-Powered Automation Software

As Caddi’s first in-house product designer after its MVP stage, I focused on transforming a complex AI-powered automation platform into a scalable, user-friendly product. My role involved designing a Chrome extension, refining the core web experience, enhancing the accuracy of AI-driven workflows through better input design, and creating a framework for onboarding and adoption. I partnered closely with engineers and ML specialists to bridge gaps between automation complexity and user clarity.

The original workflow screens lacked clarity, making it difficult for users to understand automation steps or troubleshoot failures. To address this, I redesigned the Workflow Documentation view to clearly show tasks, branching logic, and detailed steps with visual grouping. This made complex automations more digestible while maintaining the precision required by enterprise clients.

I also introduced a Run History panel with a detailed error state. Instead of generic “failed” notifications, users now see which step broke, why (e.g., invalid recipient email), and recommended actions. This design reduced dependency on support and improved the accuracy of AI-driven retry runs.

To streamline enterprise adoption, I designed an Admin Integration Hub allowing IT teams to set up and manage tools (Salesforce, Outlook, NetDocuments) at scale.