Portfolio
D2C Skincare AI Agent PILOT PROJECT
AI AGENTS · D2C SKINCARE · WHATSAPP

AI-Powered Support & Sales Assistant for D2C Skincare

A complete WhatsApp-based AI agent system that handles product recommendations by skin type & budget, answers FAQs from a knowledge base, tracks orders in real-time, and seamlessly hands off complex queries to human agents — all powered by n8n workflows and Airtable data.

8 Product SKUs
9 FAQ Entries
5 Airtable Tables
6 Agent Capabilities
n8n — Pilot WhatsApp Support Workflow ● Published
n8n Workflow Canvas showing WhatsApp trigger, normalization, handoff check, AI agent tools, and response nodes
WHAT THE AGENT DOES

Six Core Capabilities

Each capability is backed by Airtable data, ensuring the AI never invents information — every answer is grounded in verified records.

Smart Product Recommendations

Recommends products based on skin type, concerns, and budget from verified catalogue data. Supports Hinglish conversation.

FAQ Auto-Response

Instantly answers shipping, returns, COD, and delivery questions from a structured FAQ knowledge base in Airtable.

Order Tracking

Retrieves order status, items, total amount, tracking ID, and estimated delivery from the Orders Demo table.

Human Handoff

Detects complex queries, complaints, or explicit requests for human support and seamlessly routes to a live agent.

Conversation Logging

Every interaction is logged in Airtable with timestamps, intents, and responses for debugging and analytics.

Safety Guardrails

Never invents prices, discounts, or medical claims. Responds with safety disclaimers for sensitive skincare queries.

LIVE DEMO SCREENSHOTS

See the Agent in Action

Real WhatsApp conversations showing the AI agent handling product queries, FAQs, and order tracking in Hinglish.

01

Product Recommendation

Customer asks for moisturizer for oily skin within ₹1000 budget (in Hindi). Agent recommends relevant products from catalogue with prices and links.

Skin Type Detection Budget Filtering Hinglish Support
WhatsApp conversation showing AI recommending Hydrating Gel Moisturizer ₹499 and Niacinamide Serum ₹549 for oily skin
02

FAQ Auto-Response

Customer asks about shipping charges. Agent instantly retrieves the answer from FAQ knowledge base — ₹49 for orders below ₹799, free above.

Instant Response FAQ Knowledge Base Accurate Pricing
WhatsApp conversation showing AI answering shipping charges question with accurate pricing
03

Order Tracking

Customer asks "kya mera koi past order hai?" (in Hindi). Agent retrieves complete order details — ID, status, items, total, tracking, and delivery date.

Order Lookup Real-time Status Tracking ID
WhatsApp conversation showing AI retrieving past order DEMO-1001 with shipped status and tracking details
WORKFLOW ARCHITECTURE

How It Works Under the Hood

Built on n8n with color-coded workflow sections for clear separation of concerns.

n8n — D2C Pilot WhatsApp Support ● Published
Full n8n workflow canvas with color-coded sections: Orange for message handling, Blue for AI agent and tools, Red for human handoff, Green for response routing
ORANGE ZONE

Message Intake & Routing

WhatsApp Trigger → Normalize payload → Check if human handoff is needed → Route to AI or escalate.

BLUE ZONE

AI Agent & Tools

Tools Agent with access to: Catalogue, FAQ, Build Context, Model config, Orders, Dinsaagi, and Structured Output Parser.

RED ZONE

Human Handoff

When the bot detects complex/sensitive queries, routes to a human agent with full conversation context and sends notification.

GREEN ZONE

Response & Logging

Sends WhatsApp response, creates/updates Airtable log records, and stores supported conversation history.

AIRTABLE DATA MODEL

Structured Data — The Agent's Brain

Five interconnected Airtable tables serve as the single source of truth. The AI never invents data — every response is grounded in these records.

Catalogue FAQ Log Handoff Orders Demo
Airtable Catalogue table with 8 skincare products: Cleanser ₹399, Moisturizers ₹499-₹599, Serums ₹549-₹699, Sunscreen ₹749, Toner ₹449, Balm ₹249
Catalogue FAQ Log Handoff Orders Demo
Airtable FAQ table with 9 entries categorized as Shipping, Delivery time, COD, Returns, and Other with demo policy answers
📦

Catalogue

8 demo SKUs with name, category, price (₹), skin type, and key ingredients.

8 records
❓

FAQ

9 questions on shipping, returns, COD, delivery time, and safety with categorized answers.

9 records
📝

Log

Every conversation logged with timestamps, user message, bot response, and detected intent.

Live logging
🤝

Handoff

Tracks when and why conversations are escalated to human agents with full context.

Escalation records
🛒

Orders Demo

Sample orders with order ID, status, items, total, tracking ID, and estimated delivery.

Demo orders
TECHNICAL STACK

Built With

n8n

Workflow Orchestration

WhatsApp Business API

Customer Messaging

Airtable

Relational Data Model

OpenRouter / LLM

AI Response Generation

JavaScript Nodes

Custom Data Transform

Structured Output Parser

Response Formatting

LET'S CONNECT

Interested in This Project?

Reach out for a live demo, technical walkthrough, or to discuss building something similar for your D2C brand.

Dhruv Rajwada

Dhruv Rajwada

AI Automation Developer

Building AI-powered workflow systems for D2C brands. Focused on conversational AI, n8n orchestration, and structured data.

Book a Demo