01 / WhatsApp sales operations
Every WhatsApp inquiry becomes a case. AI answers first, checks stock and credit in SAP, and the sales team closes the deal from one screen.
Salam, do you have the V-belt for a Komatsu PC200?
Yes, 6 in stock at our Riyadh branch. Shall I prepare a quote?
Yes please, 2 pieces.
| Item | Qty | SAR |
|---|---|---|
| V-belt B-68, PC200 | 2 | 680.00 |
| Belt tensioner | 1 | 185.00 |
| Discount 5% | -43.25 |
QueuedReady to dispatch
02 / Overview
The sales team handled about 100 WhatsApp conversations a day across 7 branches and roughly 700 customer accounts. Smart Sales sits on top of SAP without replacing it and replaces the layer in between: personal chats, spreadsheets and email. Every message becomes a structured case, AI is the first responder, and every role gets one screen designed for its work, in English and Arabic.
03 / Problem
Customers already lived on WhatsApp, so that is where the business ran. Nothing about that flow was visible, measurable or shared.
04 / Define
The constraint that shaped every decision: customers would not install anything, learn anything or change numbers. All of the change had to happen on the staff side.
Each rep keeps their own WhatsApp Business number. Messages route into the portal and replies go out from the same number.
AI answers stock and FAQ questions instantly, day or night, and hands anything that needs judgement to a person with the full context.
Stock, credit and invoices are read from SAP. The platform adds the workflow SAP was never designed for.
05 / System
Fifteen modules sit behind one portal. A single customer message can pass through routing, AI, SAP lookups, quoting, approvals and the warehouse without anyone copying data between tools.
06 / Decisions
The thinking
Moving a case to a manager for every discount breaks ownership. The rep loses the thread and the customer feels the handover.
The decision
A discount request is a ticket attached to the case, sent to the branch manager, regional manager or management. The rep keeps the case throughout.
The result
One owner per case, and a visible approval chain with the amount and a note at every step.
The thinking
The slowest step was the credit check, and it happened in a different tool from the conversation.
The decision
Every case shows a live credit snapshot from SAP: limit, used, aging and flags, on the same screen as the chat.
The result
The answer to 'can we sell to them on credit' is on the case before the rep starts the quote.
The thinking
Staff alerts sent from a rep's customer-facing number would mix internal and customer conversations.
The decision
Customer conversations use each rep's own number. Staff notifications go out from a separate platform number, plus email and in-portal alerts.
The result
Customers only ever see their rep. Staff never miss an approval waiting on them.
07 / Solution
Journey 01
One real shape of a morning, end to end, without leaving the case.
"Do you have the V-belt for a Komatsu PC200?" arrives on Ahmed's number.
The customer is found in SAP by phone. History and credit appear beside the chat.
Stock is checked live. The customer hears back in seconds, not hours.
Ahmed builds the quote with AI-suggested lines and requests a 5% discount from his branch manager.
The approved quote PDF goes out from Ahmed's number. The warehouse sees it in their queue.
Journey 02
The same case looks different depending on who opens it.
Their cases, with AI replies already sent overnight.
Approvals waiting on them, and their team's pipeline.
Items to mark ready or dispatched, mirrored from SAP delivery.
Customers over limit or overdue, before a quote goes out.
08 / Outcome
The result is one place where every inquiry is visible, answered and owned. The business keeps WhatsApp and SAP, and loses the spreadsheets and side chats in between.
Rep, manager, warehouse, finance and leadership all work from the same case, with an audit row behind every change.
The fastest adoption comes from changing nothing the customer sees. All of the effort belongs on the staff side.
09 / Measures
The targets the platform was designed against.
First response
< 1 min
Every inquiry answered by AI, at any hour, with live stock where relevant.
Credit check
1 glance
Down from 30 to 40 minutes in a spreadsheet to a panel on the case.
Approval trail
100%
Every discount with a named approver, an amount and a note.
Next project
A Saudi AI company I co-founded and lead technically: chat and voice agents grounded in each client's own knowledge, connected to the systems they already run.