Wizabot

Wizabot is a B2B SaaS platform for e-commerce businesses: each tenant can configure WhatsApp chatbots that answer customer questions with LLM-backed AI, send template messages, and tune AI behavior and prompts from a dedicated experience.

I collaborated with AI engineers to ship this product end to end—backend services, dashboard UI, AI integration, and containerized deployment. My work was on the internal dashboard (not the public landing page), so there is no public product link or repository here.

E-commerce teams need always-on customer messaging on WhatsApp without drowning in manual replies. They need a multi-tenant product: each business configures its own bot, templates, and AI—without compromising scale, observability, or safe LLM usage.

Platform

• Self-serve configuration — businesses set up WhatsApp chatbots to respond with AI, send template messages to customers, and manage prompts and AI settings from the product.
• B2B SaaS model — built as a shared platform so many e-commerce customers run isolated configurations on the same foundations.

Backend engineering

• Express.js and TypeScript for the main API layer, with MongoDB for application data.
• WhatsApp Cloud API integration so businesses can run real channel automation at scale.
• LLM capabilities wired into the stack so chatbot replies stay useful and controllable in production.

Dashboard UI

• Designed and built the Wizabot dashboard with shadcn/ui on Next.js, Tailwind CSS, and TypeScript: a clear interface where each business configures its chatbot, manages customer interactions, and supports automated, personalized engagement—without relying on engineers for day-to-day changes.

AI integration and optimization

• Improved the Python FastAPI service that powers the AI path: better logging, cleaner structure, and solid integration with OpenAI so prompts and model calls stay maintainable as the product evolves.

Docker and deployment

• Dockerized the application using shared volumes and a Docker network so the Express app and FastAPI AI server run as a coherent system—easier to ship, reason about, and operate together.

Shipping B2B SaaS next to AI engineers means balancing product UX, API boundaries, and LLM reliability. Multi-tenant WhatsApp automation plus a separate AI service reinforced how much observability and clear contracts between services matter once customers depend on them in production.