ENGINEERING CASE STUDY / AI SAAS · MULTI-TENANT

Wasla AI.

AI WhatsApp Operating System for Businesses

A shared workspace that brings WhatsApp conversations, AI agents, business knowledge and human teams together.

MY ROLE

Architecture, frontend engineering, backend integration, AI workflows and production delivery

Next.js 16TypeScriptLaravel 13PostgreSQLpgvectorRedisMinIODockerCoolifyOpenAIEvolution API

Product overview · the organization’s operational workspace

The challenge

Businesses running support and sales on WhatsApp need more than a chatbot. They need conversation ownership, shared business knowledge, predictable automations and a clear path from AI assistance to human support, without tying their product to a single WhatsApp provider.

What I built

I designed a multi-tenant SaaS with organization-scoped data, a shared inbox, assignment workflows, AI agents, knowledge-base ingestion, RAG, deterministic automations and analytics. The frontend brings these capabilities into one operational workspace.

02 / ARCHITECTURE & TECHNICAL DECISIONS

Engineering with intent.

Organization boundaries first

Organization-scoped authorization and data flows underpin the multi-tenant architecture, keeping workspace responsibilities explicit across the frontend and API integrations.

Replaceable provider integrations

WhatsApp-specific logic sits behind a provider abstraction. Business workflows can evolve from QR/Baileys infrastructure to other providers without being rewritten.

Grounded AI, deliberate handoff

Embeddings and pgvector support retrieval from business knowledge. Structured reply and handoff decisions connect AI assistance to human ownership of conversations.

Delivery beyond the interface

Redis queues, private MinIO object storage, Docker containers and Coolify deployment support the product, alongside backup, restore and recovery practices.

03 / KEY FEATURES & ENGINEERING HIGHLIGHTS