All systems

WhatsApp inquiries that turn into orders.

One control system for the online store, from what sells to where each order stands, with an AI agent that answers customers on WhatsApp and turns inquiries into orders.

Client
A furniture retailer
Industry
Home furniture and e-commerce
Connects to
The online store, WhatsApp, invoicing, ad accounts
The system, drawn with sample data. We never show a client’s own screens.

01Overview

Most of the retailer’s customers start on WhatsApp. The control system sits above the store, the invoicing, the ads and WhatsApp. Its AI agent asks what a customer needs, checks price and stock while it answers, and passes the chat to a person when it should.

01

Shared inbox and customer page

Every conversation in one place, with claim, takeover, notes and the full customer record.

02

The control room

A switch for every capability, the AI off switch and the health of every connection.

03

The agent and its rules

A set of tools it may use, a live price and stock check, and a handover to a person.

02The challenge

Every inquiry waited for a person.

Sales and service ran through WhatsApp by hand. Customers, orders, invoices and ads sat in separate systems, so nobody had one picture of a customer. Prices and stock were quoted from memory, and the owner pieced each day together.

What the client wanted to achieve

Answer every inquiry fast

The agent replies at once and follows up on its own clock.

Quote only what is true now

Price, stock and order status come from a live check.

One view of the business

Customers, orders, invoices and ads meet in one place.

Stay in control

Every capability has its own switch, and one switch stops all AI at once.

How it worked before

  • Every chat answered by hand
  • Customer details spread across the store, chats and invoices
  • Prices and stock quoted from memory

03How we built it

  1. 01

    Mapping every system of record

    Which system owns price, stock, orders, invoices and ads, and which facts are ours to keep.

  2. 02

    Reading the real store first

    We read the store through its own interface, quirks included, before writing a line of the order link.

  3. 03

    The catalogue apart from the knowledge

    A product mirror for search, owner-approved documents for policy answers, and never a price from memory.

  4. 04

    The buying conversation, category by category

    Each category lists what the agent must know before it recommends, such as the use, the size limit and the seats for a sofa.

  5. 05

    Setting the clocks

    A gentle follow-up after ten minutes, a last one before the day closes, business hours, and a response target for chats handed to a person.

  6. 06

    The guard rails before the voice

    The agent acts only through its registered tools, and our code refuses any price or link not checked in that same turn.

  7. 07

    Rehearsed in scripted conversations

    Thirty-six conversations, from the easy sale to the refused discount, before any customer sees it.

04How it works

How an inquiry becomes an orderA WhatsApp message becomes a paid order, with a person ready to take over at any step.
  1. Step 01Inquiry

    A customer writesA WhatsApp message opens a customer record.
  2. Step 02Qualify

    One question at a timeThe agent asks only for what it still lacks.
    Nothing asked twiceEvery fact the customer gives is kept.
  3. Step 03Recommend

    A live checkPrice and stock read from the store before they are said.
    Or a personA request, a discount question or an outage moves the chat to the team.
  4. Step 04Purchase

    Pays in the storeA paid order closes the lead and stops the follow-ups.
    ConfirmationThe order is confirmed on WhatsApp.

The rules the agent keeps

  • It never states a price or stock it did not check in that turn
  • It never sends a link it was not given
  • It never offers a discount the store has not published
  • It never sees margin or cost
  • It stops at once when a customer says stop
  • It hands over to a person when the model is down
  • It suggests extras only after clear interest, never during a complaint

Everything at a glance

What is running, what is waiting and what needs a person.

A switch for every capability

And one AI off switch that stops it all at once.

How it speaks and spends

Its settings and a spending ceiling.

Health of every link

The store, invoicing, ads and WhatsApp.

05The outcome

Every conversation lands in one inbox. The agent answers at once, asks one thing at a time, checks price and stock before it says them, and hands the chat over with everything a person needs to take it from there.

BeforeAfter
Every chat answered by handAn AI agent answers, and a person can step in at any moment
Customer details spread across the store, chats and invoicesOne page per customer
Prices and stock quoted from memoryChecked live before they are said
The owner pieces the day togetherA morning brief of what sold and what needs a person

06Key figures

36scripted conversations rehearsed before launch
12tools the agent can use, and nothing else
50checks in the acceptance list

07Technology

What the system runs on, for the technical reader.

Cloud
AWSContainersManaged PostgreSQLAlerts and monitoring
AI
Claude
Messaging
WhatsApp Business Platform
Integrations
The online store platformThe invoicing serviceAd accounts
How we build
TypeScriptNext.jsTest-first, with a stand-in for every connected system

08What comes next

Next, WhatsApp and the store connected for real, then a careful launch, one step at a time.

The launch path

  1. 01

    A shadow run

    The agent drafts every reply and the team sends it.

  2. 02

    The owner’s sign-off

    Nothing goes further until the owner agrees.

  3. 03

    A quarter of new conversations

    The agent answers some customers on its own.

  4. 04

    Every conversation

    After three clean business days.

  5. 05

    Two weeks of close watch

    Every stop still works throughout.

09About

About the client

An online furniture retailer whose customers mostly start on WhatsApp. We name a client only with their permission.

About MYBB

MYBB builds systems around how a business already works. We diagnose it with MYBB ROIQ, build the system, launch it with the team and keep running and improving it.

All systems

Questions about our work.

Can you build something like this for my business?

Each system here was built around one business. Answer the 10 questions, and within 48 hours your growth plan shows what we would build for yours.

Why don’t you name the clients?

We name a client only with their permission. Until then, each system is described by the kind of business it serves.

Are these the real screens?

They are redrawn with sample names and numbers. We never show a client’s own screens or data.

My business is in a different industry. Does that matter?

No. We start from how your business works, not from a template for its industry.

Will we have to replace the tools we use?

We start with what already works. The system may keep some of your tools, connect them, and add what’s missing.

Who decides what the AI does?

We agree with you where AI helps and where a person reviews, decides or approves. Those rules are part of the blueprint.

Who runs the system once it’s built?

We do, with your team. We keep running it and improving it as your business changes.

How is the price set?

By the scope, the connections and what the rollout needs. The price is in the blueprint, and nothing is built before you approve it.

AI agents that turn questions into orders.

We build agents that answer your customers and take them to an order, with your team in control of every step. Tell us how your store sells.

Prefer to talk first? Email welcome@mybb.ai