home / use cases / messaging channels / whatsapp-customer-service
Messaging channels On the roadmap

Customer service where customers already are

A customer messages your WhatsApp business number and your Aivell answers — hours, stock, prices from your knowledge base, order status from your database, in the customer's language, within a minute. Complaints and money questions go to a human. Reading and drafting happen on the box in your office.

Watch the demo

starts when: A customer messages your business number on WhatsApp

The problem

Your customers are not composing emails. They are sending a WhatsApp message from the bus — “is the red one in stock?”, “where’s my order?” — and judging you by what happens in the next ten minutes. Answering within a minute lifts conversion by 45% or more; answering at closing time tomorrow mostly lifts your competitor’s.

For a small business the channel is a trap: the phone that owns the number goes home with someone, evenings and Saturdays pile up unanswered, and the same five questions arrive daily. The tools that promise to fix this route every customer conversation through a chatbot vendor’s cloud.

How your Aivell does it

This channel is rolling out to early-access customers first — here is what it does when it reaches you:

  1. The message reaches the box. WhatsApp delivers it; from there, everything — reading, looking up, drafting — happens onboard, behind your firewall.
  2. The everyday questions get instant answers. Opening hours, prices, stock, the returns policy: answered from your own knowledge base, in the customer’s language, with your tone. No source, no answer — the chore hands over instead.
  3. “Where’s my order?” gets a real answer. The Aivell finds the order in your database by the customer’s number and replies with the actual status. The database guardrail is read-only: it can look, never change.
  4. The delicate ones go to Giulia. A complaint gets an apology and a human, not a chirpy reply. Discounts and refunds are forbidden territory by rule — the Aivell can’t promise what you didn’t approve.
  5. Nothing happens off the record. Every conversation and outcome lands in the daily sheet, so Monday’s coffee comes with a view of what the weekend asked.

The number stays yours, the answers come from your documents and your systems, and the thinking happens in your office — which is a strange thing to be able to say about a WhatsApp chatbot, and exactly the point.

The chore, as you'd write it.

No flowcharts, no code — a chore is just a message to your Aivell, in your own words. It shapes it into a solid, guarded procedure and follows it to the letter. This one:

Giulia Aivell · answer whatsapp customers ● active

greet within a minute, in the customer's language

answer hours, prices and stock from the knowledge base

order status? look it up in the database by the customer's number

complaints get an apology and Giulia — not an answer

never promise a discount or a refund; money is for humans

after 21:00, answer what you can and queue the rest for morning

still stuck after two tries? say so honestly and hand over

log every conversation and its outcome in the daily sheet

You
Got it — I'll start right away, and check with you whenever something needs a human.
Giulia Aivell

The tools it uses.

Connected with single-click passkeys — no copied tokens, no OAuth hell.

WhatsApp reads and replies on your business number
Knowledge Base answers hours, prices, policies
Database looks up orders, read-only
Email hands the hard ones to Giulia

Sensible guardrails.

Every tool can be limited to exactly what this chore needs. For this one, you might set:

WhatsApp Replies only to customers who wrote first. No outbound campaigns, no bulk messages, ever.
Database Read-only. Order lookups are allowed; changing an order, a price or a customer record is blocked.
Knowledge Base Answers come from the approved folder; when no source covers the question, the chore says hand over, not improvise.

Questions, answered.

When can we have this? +

WhatsApp support is rolling out to early-access customers first. Tell us it matters to you and we'll put you on the list — everything else on this page is how it's built to work.

Do customer chats end up on an AI provider's servers? +

No. WhatsApp delivers the message as it always does; the reading, the database lookup and the drafted reply all happen on the box in your office. No conversation is sent to a cloud AI to be understood.

Will customers think they're talking to a human? +

They shouldn't, and the chore makes sure of it — it answers plainly as the shop's assistant and hands over honestly when it's out of its depth. Fast and straight beats a fake first name.

What is it never allowed to do? +

The guardrails are blunt — it cannot message anyone who didn't write first, cannot change anything in your systems, and the chore forbids promising discounts or refunds. The worst case is an answer deferred to a human, not a commitment you didn't make.

Related.

What is an Aivell?

An Aivell is an on-premise AI colleague: a small box that plugs into your network, runs its own AI onboard and takes the repetitive work off your desk. Prompts, documents, data — nothing ever leaves your office. Unplug the internet: it keeps working.

You describe each task as a chore, in plain language. Your Aivell turns it into a solid, guarded procedure and runs it in the background — with guardrails on every tool, a complete audit trail, and approvals in your hand when you want them. One-time setup, fixed monthly fee. No tokens, no overages.

Watch the demo

onboard inference · air-gapped · fixed monthly fee