Work / AI agents / Live

WhatsApp customer support agent for a D2C brand

An AI agent on WhatsApp that answers order, delivery and return questions using live store data, and hands complex cases to the support team.

Problem

Most support messages asked the same things — where is my order, how do I return this, is this in stock — and customers waited hours for a reply during sales.

Objective

Resolve routine questions instantly on WhatsApp with accurate, order-specific answers, and route everything else to a person with full context.

Solution

  1. 01

    The WhatsApp Business API delivers messages to an agent service that keeps conversation state per customer.

  2. 02

    The LLM uses narrow tools — look up order, check shipment, start return, check stock — backed by the store and courier APIs.

  3. 03

    Customers are verified by phone number and order ID before any order details are shared.

  4. 04

    Low-confidence answers, complaints and refunds above a limit go to the helpdesk with a summary of the conversation.

Architecture

01
Customer on WhatsApp
02
WhatsApp Business API
03
Agent service
04
LLM (tool use)
05
Store + courier APIs
06
Helpdesk hand-off

AI workflow

01

Receive the message and load the customer's conversation state

02

Verify the customer before sharing order details

03

Call the relevant tool: order, shipment, return or stock

04

Answer in the customer's language with the live data

05

Hand off to a person when confidence is low or policy requires it

Technology & role

WhatsApp Business APIClaude APIPython / FastAPIPostgreSQLShopify API
Our role

Conversation design, agent and tool development, integrations, hand-off workflow and deployment.

Start here

Want something similar built?

Send a short brief. We'll reply personally with questions or a suggested next step — usually a technical discovery call.