Airbox helps sales teams manage incoming leads from various social media sources, track their journey through a pipeline, and convert them efficiently. As the sole product designer on this project, I was responsible for the full design system: brand guidelines, the web app, and now the mobile app, all from scratch.
Problem
Real estate sales teams operate across multiple stages — from the moment a lead comes in, to site visits, negotiations, and final booking. Each stage has different team owners, different time sensitivities, and different conditions that have to be met before a lead moves forward.
The client needed a way for admins to build and configure these pipelines without requiring developer involvement.
The design challenge: How do you make pipeline creation feel intuitive for a non-technical admin — while still being flexible enough to handle complex sales workflows?
Creating a canvas pipeline builder
My initial solution was a canvas-based pipeline builder where admins could visually add stages, assign teams, set conditions (mandatory notes, Q&A prompts), configure reminders, and define reallocation timelines.
Enhancing the experience through AI
I proposed to use the global pipeline as a structured JSON reference that Box AI works from. The admin describes their intent in plain language. AI maps that input onto the existing template — it doesn't generate pipelines from scratch, it configures a known structure. This preserves data integrity on the backend while keeping the frontend experience genuinely conversational.
What does Box AI solve?
Impact
We expect pipeline setup time to drop from ~20 minutes of manual configuration to under 3 minutes with Box AI.
