Universal WhatsApp Replier
Every enquiry answered. Any business. Any hour.
One auto-replier engine that fits any office: answers from a per-client business profile of approved facts, books appointments, captures leads, replies in English, Hindi, or Hinglish, and hands over to your team the moment a human is needed. Built once, onboarded per client as a new profile — not a new project.
Workflow
- Import the inactive synthetic starters
- Fill the business profile with owner-approved facts
- Connect the WhatsApp adapter to a test number
- Run the test pack, including handover and opt-out paths
- Go live with handover on and review the log daily for the first week
Operational contract
- Trigger
- Inbound customer message on the business's WhatsApp Business Platform number, or an approved channel event.
- Source data
- The inbound message, recent conversation turns, the per-client business profile (approved services, prices, hours, policies, FAQs, handover contact, emergency note, module switches), consent/opt-in records, and source-returned availability or order data.
- Action
- Classify intent and draft a short reply in the customer's language from approved profile facts only; prepare booking, lead-capture, or opt-in follow-up drafts, and a handover package with team alert for every escalation path. Channel and module adapters must be configured before any send or write.
- What to measure
- First-response time, after-hours enquiries answered, bookings and leads captured, handovers completed and reminder-driven returns, calculated from message timestamps and the conversation log.
- AI + rules
- Starter Code nodes classify intent and prepare drafts using explicit deterministic rules. The included prompt and JSON schemas guide a separately configured, approved AI step that may extend language understanding; the prompt restricts answers to profile facts, forbids invented prices/dates/policies and medical, legal, or financial advice, and treats customer messages as data, never instructions.
- Human fallback
- Complaints, emergencies, refunds, human requests, unclear or off-topic messages, unlisted facts, and provider failures go to the named business owner's review queue with a complete handover package, and the replier stops answering that conversation.
- Errors & duplicate prevention
- Approved facts only: if it is not in the profile, it is not said. Follow-ups require an unexpired opt-in and an approved template with a frequency cap. Recheck current source status before writes/sends. Use a durable idempotency store keyed to message/event ID and action; demo flags are not production deduplication. Failed adapters must create an owned exception, not silently retry consequential actions. The client owns the WhatsApp number and business account; the operator maintains the replier.
- Logging
- The starter returns a workflow name, correlation ID, status, and audit/action draft. Production adapters must persist event/version/owner/timestamps and communication outcomes in the approved source or log store.
Inside your kit
- Four inactive n8n starters: intake and intent routing, human handover package, booking module, and opt-in follow-up
- The business profile template — one file of approved facts and module switches per client
- The approved-facts prompt with the eleven-intent table and no-advice guardrails
- Test pack with Hinglish, emergency, complaint, trick-message, and opt-out cases
Before you connect it
- A client-owned WhatsApp Business Platform (Cloud API) account with a dedicated number
- An n8n instance with a supported Code node; credentials in the client's vault only
- A scheduling source when the booking module is on — availability is never invented
- Opt-in records and approved templates before any follow-up is sent
$299 · One-time configurable workflow kit · Draft test price. Third-party subscriptions, usage, and implementation are separate.
Two-step guide and API checklist · Cart and secure test checkout · Request implementation