← Back to Work
7UNIT / CASE STUDY 002
SMB Sales · UAE · 2025

WhatsApp-native sales CRM forteams that close inside conversations.

THE PROBLEM

UAE sales teams were managing their entire client relationship inside WhatsApp — but their CRM was a separate tool that required manual data entry after every conversation. Deals were falling through the gap between where conversations happened and where data lived. Pipeline visibility was zero unless the salesperson manually updated the CRM after every message.

WHAT WE BUILT

A CRM system that operates entirely within WhatsApp. Pipeline tracking, AI-triggered follow-up automation, lead qualification, and team assignment — all without leaving the conversation. The system reads conversation context, maintains deal state, and surfaces the next action without requiring the salesperson to touch a separate tool.

ARCHITECTURE DECISIONS

WhatsApp as the primary interface

The temptation was to build a web portal with WhatsApp as a notification channel. We rejected this. The product had to live inside WhatsApp itself — because that is where the salesperson's attention already is. Every feature was evaluated against one question: does this require leaving WhatsApp?

Rationale: UAE sales teams do not switch tools mid-conversation. If the CRM requires a context switch, it will not be used. The product had to live where the work happens.

Trade-off accepted: WhatsApp API limitations constrain some UI patterns — accepted in exchange for zero context-switching for the end user.

AI-triggered follow-up

Silence detection with configurable threshold. When a deal has been inactive for N days, the system sends a personalised follow-up — not a template blast. The AI generates context-aware messages based on the last conversation thread.

Rationale: Generic follow-up templates have low response rates. A message that references the specific conversation context performs significantly better. The AI generates this from the conversation history rather than from a static template.

Trade-off accepted: Slightly higher per-message cost from AI generation in exchange for meaningfully higher response rates.

Session memory across conversations

Conversation context preserved across sessions. The salesperson can pick up a conversation after two weeks and the system knows the deal stage, the last touchpoint, and the next suggested action.

Rationale: Deals in UAE often span weeks or months. A CRM that loses context between sessions forces the salesperson to re-read entire conversation histories. Session memory eliminates this.

Trade-off accepted: Higher storage requirements for conversation state in exchange for seamless continuity across sessions.

OUTCOMES
  • Pipeline managed entirely within WhatsApp — zero CRM context-switching
  • AED 85,000 stalled deal recovered by AI follow-up
  • Follow-up response time: 38 minutes average after AI trigger
STACK
WhatsApp Business APIFastAPIPostgreSQLClaude APIReact
CONTINUE EXPLORING

Related case studies

EdTech · Italy · 2024

Holistic growth centre portal with

How we built a learner portal connected to Odoo with event-driven curriculum sync and automated certificate generation for 1,200 learners in Italy.

Read case study →
AI Systems · Global · 2025

LLM-powered assistants built for

How we build LLM-powered assistant systems designed around real production constraints — session memory, human handoff, cost attribution, and tenant isolation.

Read case study →