Collabo

AI-powered influencer campaign manager for D2C brands. Forward WhatsApp chats or screenshots → Gemini extracts deals, tracks deadlines, sends reminders, and manages proofs automatically.

Comment

Inspiration

I noticed that many small D2C brands and boutique agencies in India manage dozens of micro-influencers through messy WhatsApp chats, screenshots, and Google Sheets. They are great at marketing but lose track of deliverables, deadlines, and payments. A single missed deadline or forgotten payment can damage both the brand and the influencer relationship.

I wanted to build a tool that turns the chaos of WhatsApp negotiations into clean, trackable campaigns using AI — without forcing brand owners to learn a complicated new system.

What I Built

Collabo is an AI-powered influencer campaign management platform designed specifically for D2C brands and small agencies.

Key features:

  • Forward a WhatsApp chat, screenshot, PDF, or voice note → Gemini AI extracts influencer name, deliverables, deadline, and payment amount
  • Automatic campaign creation and tracking dashboard
  • Automated email + WhatsApp reminders 48 hours before deadlines and when overdue
  • Magic-link proof upload so influencers can submit content without creating an account
  • Status workflow (Draft → Active → Content Received → Needs Revision → Approved → Paid)
  • Simple analytics and link tracking

The goal was to make the product feel like a natural extension of WhatsApp rather than another heavy CRM.

How I Built It

  • Frontend: Next.js 15, Tailwind CSS, shadcn/ui
  • Backend: FastAPI (Python)
  • Database & Auth: Supabase (with Row Level Security)
  • AI: Google Gemini (multimodal extraction from images, PDFs, and voice notes)
  • Messaging: Meta WhatsApp Cloud API + Telegram
  • Payments: Razorpay
  • Hosting: Vercel (frontend) + Render (backend)

I focused heavily on reliability — graceful AI fallbacks, rate limiting, Redis locks for background jobs, and clean error handling so the product never crashes for the user.

Challenges I Faced

  1. AI Extraction reliability – Real WhatsApp screenshots and voice notes are messy (Hinglish, partial information, low-quality images). I had to build multi-stage fallbacks and human-in-the-loop review.
  2. WhatsApp multi-user bot – Making one bot number work for many brands while keeping data isolated required careful phone-number linking and webhook security.
  3. Background jobs on free/cheap hosting – Ensuring reminders fire reliably without duplicate messages when the server restarts or scales.
  4. Keeping the product simple – Many features were tempting, but I constantly cut scope so a busy brand owner could start using it in under 5 minutes.

What I Learned

Building a real SaaS taught me that technical complexity is only half the battle. The harder part is making the product feel effortless for non-technical users. Every time I added a new feature, I asked: “Would a tired brand owner actually use this after a long day?”

I also learned the importance of graceful degradation — when the AI fails, the user should still be able to complete the task manually without frustration.

Collabo is still early, but it already solves a painful daily problem for the people I built it for.

Built With

Share this project:

Updates