AI / CRM / SaaS
AI Lead Inbox
AI-assisted lead operations platform that centralizes intake, qualification, pipeline movement, and reply workflows.
Role & Contribution
Project creator · Product design · UI/UX design · Full-stack development
Built the product independently, from the initial idea and user flows to interface design, frontend, backend, and integrations.
Platform
Multi-surface web platform with workspace operations, public intake forms, kanban workflow, and lead communication flows.
Stack
React 19, Vite, Tailwind CSS, TanStack Query, NestJS, TypeORM, PostgreSQL, BullMQ, OpenAI API
Interface Gallery
The gallery walks through the full workspace flow: main dashboard, projects, project-level controls, lead handling, the AI profile, public intake, and both sides of the chat experience.
Main workspace dashboard
Projects workspace
Project overview and analytics
Channels and intake setup
Notification logic
AI profile: business context
AI profile: qualification and replies
Project settings and intake key
Next-step queue
Lead board with statuses
Opened lead profile
Public intake form
Customer-side lead chat
Owner or staff reply chat
Overview
AI Lead Inbox is a product system for teams that need one operating layer for incoming demand. It unifies public intake, channel-based lead capture, qualification, follow-up, and owner control inside one clear workspace.
Challenge
The core challenge was to turn scattered inbound requests into a disciplined operating process. Leads arrive from different channels, require fast qualification, and quickly become expensive to manage when the review and reply flow is fragmented.
Solution
I designed the case around a complete lead loop: intake enters one inbox, AI helps classify and prioritize, the team works through review and kanban states, and the conversation continues through internal and client-facing reply surfaces.
Product Flow
A reusable flow map shows how users or data move through the product from entry to outcome.
Lead Source
Inbox
AI / Classification
Review
Pipeline / Kanban
Follow-up
Key Features
Lead inbox
Kanban workflow
Filters and search
Lead detail view
Dashboard
Onboarding
Responsive mobile workflow
Architecture / Technical Overview
This diagram stays intentionally high level and avoids exposing sensitive implementation details.
Lead Sources
Web Application
Business Logic
Data Layer / Integrations
Technical Decisions
- • Separate the frontend and backend into distinct workspaces so UI delivery and operational logic can evolve independently
- • Use NestJS with modular domains and TypeORM persistence to keep lead, auth, and workflow logic structured
- • Support async follow-up, queue-based work, and AI calls through BullMQ instead of pushing everything into request-response paths
Challenges
- • Keeping owner controls dense enough for daily operations without making the interface heavy
- • Balancing AI-assisted triage with a workflow that still feels predictable and editable by the team
- • Connecting intake, pipeline, deadlines, and replies into one loop instead of separate disconnected screens
Result
The result is a coherent lead-operations product: teams can capture demand, qualify it, assign responsibility, move it across statuses, and continue the conversation from one responsive environment.
Next project
Evidra
Document processing workspace that turns invoices and scans into structured data, with field-level review, team comments, and approved exports.