CO-FOUNDED VENTURE

Dibs

Investor-backed AI Omnichannel CRM orchestrating autonomous conversation agents for service businesses across Southeast Asia.

usedibs.com

The Thesis & The Problem

The problem came from a direct human insight: service businesses—such as salons, studios, and clinics—leak revenue in two distinct ways. First, popular timeslots sell out while off-peak slots sit empty, creating inefficiencies in capacity. Second, customer enquiries pour in through WhatsApp, Telegram, and Instagram simultaneously, with many falling through the cracks because there is no unified system to manage them.

The Mindset: We didn't start with "let's build an AI tool." We started with "how do service businesses actually lose money?" and designed backwards from there. Technology must always be in service of a human problem.

Omnichannel Architecture & Dynamic Pricing

I designed Dibs as an AI-powered omnichannel CRM to directly target these revenue leaks. The platform aggregates all messaging channels into a single unified inbox, powered by an AI agent that converses with customers naturally to convert enquiries into bookings, create rich customer profiles, and re-engage idle leads on autopilot.

Beyond simple conversational AI, I engineered a dynamic pricing layer. This engine analyzes historical booking patterns to automatically price popular slots higher and intelligently discount underperforming ones—effectively stopping capacity-driven revenue leakage at the source.

System Architecture & Tech Stack

As the Technical Co-Founder, I architected the entire application from the ground up to support high-volume, multi-tenant B2B operations:

Frontend: Architected a high-performance React dashboard to serve as the unified inbox for merchants, built to handle real-time omnichannel syncing with zero UI blocking.
Database & Multi-Tenancy: Designed scalable, highly relational Supabase PostgreSQL schemas to strictly isolate multi-tenant data while supporting complex AI context retrieval operations.
Billing Infrastructure: Integrated Stripe to power complex tiered SaaS subscriptions and dynamic booking charges.

Core Database Schema

Below is a high-level representation of our relational architecture that maps omnichannel conversations and business configurations securely:

-- Relational Schema for Omnichannel Messaging layer CREATE TABLE businesses ( id UUID PRIMARY KEY DEFAULT gen_random_uuid(), name VARCHAR(255) NOT NULL, stripe_customer_id VARCHAR(255), created_at TIMESTAMP WITH TIME ZONE DEFAULT now() ); CREATE TABLE omnichannel_conversations ( id UUID PRIMARY KEY DEFAULT gen_random_uuid(), business_id UUID REFERENCES businesses(id) ON DELETE CASCADE, channel VARCHAR(50) NOT NULL, -- 'whatsapp' | 'telegram' | 'instagram' external_user_id VARCHAR(255) NOT NULL,-- Twilio or Telegram user identifier session_status VARCHAR(50) DEFAULT 'active' ); CREATE TABLE messages ( id UUID PRIMARY KEY DEFAULT gen_random_uuid(), conversation_id UUID REFERENCES omnichannel_conversations(id) ON DELETE CASCADE, direction VARCHAR(20) NOT NULL, -- 'inbound' | 'outbound' content TEXT NOT NULL, is_ai_drafted BOOLEAN DEFAULT false, created_at TIMESTAMP WITH TIME ZONE DEFAULT now() );

Business Validation

The technical execution directly unlocked business traction. The robust Stripe architecture led directly to our acceptance into Stripe for Startups. More importantly, we secured institutional backing and funding from Agile & Sandbox, validating the platform's viability and setting the stage for serving paying customers across Singapore, Thailand, and Indonesia.