Skip to content
Conversational Commerce · n8n

ChatCart

WhatsApp storefront on n8n — customers browse a Multi-Product catalog and check out with WhatsApp's native cart, while the workflow enforces live stock from Sheets, stages orders, and handles FAQs and cancellations in chat.

    n8nWhatsApp Cloud APIMulti-Product TemplatesGoogle Sheetsn8n Data TablesMeta Webhooks

The Problem

Small businesses sell on WhatsApp but run it manually: forwarding product photos, confirming prices by hand, losing track of who ordered what, and overselling items whose stock they forgot to update — every order costs the owner attention.

The Solution

A full storefront living inside a WhatsApp thread, on the official Cloud API. One n8n webhook answers Meta's verification handshake and receives all events; messages route by type into interactive replies, native cart orders, or text. Customers browse via Multi-Product Templates and pay-checkout-style cart flows; each placed order is staged in a data table keyed by phone and unique order ID, validated against live inventory in Google Sheets with per-order limits, and confirmed instantly — or bounced with an out-of-stock reply. Welcome flows, an FAQ responder, and a cancellation handler cover the rest, while a compliance branch pushes business and grievance-officer details for India-market requirements.

How the workflow runs

  1. 01 · Verify

    The webhook completes Meta's subscription handshake itself and filters inbound noise before any logic.

  2. 02 · Route

    Every event is classified as interactive reply, native cart order, or plain text and sent down its own branch.

  3. 03 · Browse

    A Multi-Product Template renders the store inside chat — scroll, inspect, add to cart natively.

  4. 04 · Stage

    Placed carts land in an orders data table with phone, name, and a unique order id for fulfilment tooling.

  5. 05 · Validate

    Each line item checks against live stock in Google Sheets under per-order limits; success or out-of-stock replies fire immediately.

  6. 06 · Support

    Text senders get the welcome menu, FAQ answers, or a cancellation confirmation without human intervention.

AI layer

Native commerce UX
WhatsApp's own catalog and cart primitives do the heavy lifting — no links out, no webviews to break.
Deterministic routing
Message-type classification drives every branch, keeping bot behaviour predictable at scale.

Automation layer

Self-verifying webhook
Verification handshake handled in-flow, eliminating a separate verify endpoint.
Sheets-backed inventory
Stock lives where shop owners already update it — the bot reads it live per order.
Staged order pipeline
Data-table staging with unique IDs makes every order addressable by downstream fulfilment automation.
Compliance automation
Business entity and grievance-officer details push to the API programmatically for Indian market rules.

Technical challenges

  • WhatsApp commerce breaks when stock lives in a spreadsheet nobody syncs — customers order items that are gone.

    Per-item availability checks against Sheets at order time, with explicit limits and instant out-of-stock messaging.

  • Meta's webhook requires a verification handshake plus strict payload filtering before any business logic can run safely.

    A single self-answering webhook node handles GET verification, while a filter gate drops malformed events early.

  • Orders arriving as chat messages are ephemeral — nothing downstream can reference them reliably.

    Every order is staged with a phone-keyed unique ID in a data table, turning chat events into addressable records.

Outcome

  • A complete shopping experience inside WhatsApp: browse, cart, order, confirmation — zero human handling.
  • Overselling eliminated: stock validation runs on every order before any confirmation is sent.
  • Owners keep managing inventory in Sheets; the bot does everything else.

Lessons learned

  • Use platform-native commerce primitives (catalogs, carts) instead of rebuilding shopping UX in text.
  • Treat chat events as records: staging with unique IDs turns conversations into systems.
  • Compliance endpoints belong in automation too — one workflow node beats a manual portal form.