IdeaBoxAI helps teams bring AI into their everyday work. It brings together AI personas, reusable skills, company knowledge, and automations in one workspace, with tools for administrators to manage access, usage, and billing.
Visit ideaboxai.com ↗For much of my time at IdeaBoxAI, I was the sole Product Designer, working with the founders, product manager, and engineers to take the product from early concepts to a live platform.
My work covered both the experiences people use to work with AI and the controls organizations need to manage it.
A conversational workspace with suggested prompts and a composer for working with AI.
Discover and manage AI personas tailored to different business roles.
Browse reusable AI skills and add them to the workspace.
Organize documents and structured data that ground AI responses in company knowledge.
Track automation runs, review outputs, and resolve issues.
Introduce workspace users to their role-specific personas and guide them through setup.
Monitor AI credit usage and concurrent user capacity, review requests, and configure limits.
Manage available models, company-owned credentials, and access for roles and individual users.
I’ll use Automation and Knowledge base concepts to share how I identify problems, explore design directions, and refine solutions through iteration.
Concept explorations
Upcoming · Est. Fri, Oct 16As features expanded and more interfaces were produced with AI tools, patterns began to drift. I developed and maintained a Figma design system to bring greater consistency to the platform.
The library covered typography, colors, spacing, the app shell, and reusable components. Alongside it, I redesigned shared patterns for feature landing pages, cards and lists, search, filters, and sorting.
These components also became the foundation for AI-assisted page assembly. I documented interaction rules and worked with engineers to support gradual implementation across the live product.
Click a card to explore the documentation, including variants, states, and usage guidance.
As AI tools developed, my design workflow evolved too. I moved from designing flows and interfaces in Figma to exploring ideas through rapid AI-assisted prototypes. After exploring the strengths and limitations of each approach, I combined these tools with a shared design system into a workflow that helped me work more efficiently while maintaining design quality.
I translated requirements into rough layouts and wireframes, then developed high-fidelity screens and interactive prototypes in Figma. These became the basis for design discussions and developer handoff.
Requirements → Wireframes → Validation → Hi-fi prototypes → Implementation → Testing
The whole team began using AI tools such as Lovable and Claude to turn ideas into working prototypes. This helped us review concepts quickly, gather feedback, and enable engineers to implement new features faster under tight deadlines.
Requirements → Working prototype → Validation → Implementation → Testing
AI tools helped us prototype and develop features faster, but as the platform grew, inconsistencies in copy, spacing, components, and interactions became more apparent. We carried out a UI/UX revamp and refined our workflow, keeping Figma as the single source of truth for development.
We continued using AI tools to explore ideas through working prototypes, then rebuilt the screens in Figma using our component library. With assistance from Claude, this process became much faster. I then reviewed the user experience and refined the details before handoff, creating new components where needed. This made precise adjustments easier and reduced repeated edits that consumed credits in AI tools.
Requirements → Working prototype → Validation → Library-based assembly → Design refinement → Implementation → Testing