Case study / Software · AI · Automation
A configurable WhatsApp assistant for service businesses. It guides customer intake, answers approved FAQs, prepares supported quotes from business-approved data, coordinates eligible appointments, and safely redirects special requests to your team with the customer details and conversation history ready.
System overview
WhatsApp AI Assistant
Conversations
Guided intake
Approved answers
Business data
Scheduling
Appointments
Team support
Customer details ready
How the assistant organizes each conversation · Conceptual preview
01
Overview
This project is a persistent, always-on WhatsApp customer-service assistant. It guides customers from their first message through service intake, supported questions, and appointment coordination — then connects them with a team member at the right moment, carrying forward the details already collected and the conversation history without interrupting automation for every other conversation. It is built as one reusable engine configured per client, with separate business data, credentials, messaging session, and database for each company.
02
Context
Customers describe service needs in natural language, while the business needs consistent intake, approved information, dependable scheduling, and a simple way for staff to join a conversation with the customer information ready. The system connects those needs without treating AI as the source of business facts.
03
Solution
AI classifies intent and extracts customer-provided details. Deterministic Python services decide the next step and control authoritative information such as prices, policies, service areas, schedules, and booking eligibility. Unsupported requests are safely redirected to a member of your team, with the customer details already collected and ready for follow-up.
04
Capabilities
05
Architecture overview
A public, non-proprietary view of how a message moves through the system — from an incoming WhatsApp message to a delivered response.
Customer message in
Messaging bridge
Linked-device routing
FastAPI app
State machine & services
AI + rules
AI understands; rules decide
Data & scheduling
SQLite + Zoho Bookings
Response
Reply delivered
Customer message in
Messaging bridge
Linked-device routing
FastAPI app
State machine & services
AI + rules
AI understands; rules decide
Data & scheduling
SQLite + Zoho Bookings
Response
Reply delivered
06
Engineering considerations
The assistant is deliberately hybrid rather than free-form: AI understands language, but validated application rules decide business behavior. Conversation state is persisted, duplicate delivery is guarded, and special requests move smoothly to a team member with the collected customer details ready. Private operations, credentials, customer data, and internal access details remain outside this public portfolio.
07
Current status
The core bot, structured language-understanding flow, scheduling integration, persistence, team-routing logic, and private operator tooling are implemented, with automated validation covering the core application. Production WhatsApp messaging remains intentionally inactive while final business data is approved, the company phone is paired, and controlled live coexistence testing is completed.
This case study