Conversational Commerce: Turning WhatsApp and Social Chats Into Tracked, Profitable Orders

At 11:40 on a Thursday night, a shop owner's phone shows 63 unread chats after an Instagram story went well. In one of them, a customer asked about the grey lamp at 9:10, was told "yes, available" from memory, and sent a payment screenshot at 9:12.

The lamp was sold at the counter at 4 pm. The screenshot shows 48.00 and the price is 54.00. The address is inside a voice note, and the thread is already buried under forty others. This is an illustrative scene, not a real shop, but every detail in it is common.

The sale was not lost to chat. It was lost because chat was asked to do the job of an order system. This article shows how to keep the speed of chat and add the discipline of a proper order system: a clean flow from enquiry to delivery, the current WhatsApp Business Platform rules (August 2026), where AI helps and where a person must take over, the numbers to watch, and the mistakes that cost the most.

Why chat became a sales channel

Chat wins because it removes steps. There is no account to create, no page to load, no form to fill. The customer already trusts the app, already has it open, and can ask a human question and get a human answer in minutes.

It also fits how people actually decide. Shoppers want to know whether the fabric is thick, whether the shade matches the photo, whether delivery reaches their district by Friday. A product page answers some of that. A conversation answers all of it, which is why a fast answer can beat a lower price.

The same strength creates the weakness. A chat thread is built for talking, not for recording. It has no stock level, no payment status and no order number, so every sale made there is only as reliable as the memory of the person who made it.

The hidden cost of orders that live only in chat

Teams rarely measure what unstructured chat selling costs, because the losses are scattered. They show up as small, repeated problems:

• No stock check. The seller says "yes, available" from memory. The item was sold at the counter an hour ago, and the customer gets an apology instead of a parcel.
• Payment by screenshot. Someone has to compare each image with the bank or wallet statement. Edited screenshots exist, and honest mistakes (wrong amount, wrong reference) are easy to make.
• Lost follow-ups. A customer who says "I'll confirm tomorrow" is never contacted again, because nothing reminds anyone.
• Double handling. The same address and phone number are typed into a spreadsheet, then a courier portal, then an invoice.
• No data. You cannot say how many chats became orders, how long replies take, or which product causes the most questions.

Each one is small. Together they cap how far a chat-led business can grow, because every extra order needs extra manual work.

A clean flow from enquiry to delivery

The fix is not to leave chat. It is to make chat the front door of a system that records everything behind it. The conversation stays human and fast. The order, stock and payment steps happen in software.

In practice, the flow has six steps:

1. Enquiry. The customer writes from an ad, a post, a website button or a QR code. The conversation is linked to a customer record from the first message.
2. Product card. Instead of typing descriptions, the seller (or an assistant) sends a catalogue item with photo, price and variants taken from the live catalogue.
3. Order capture. Items, quantity, address and phone go into a real order, not a note. Stock is checked and reserved at that moment.
4. Payment. The customer receives a payment link, or the order is marked cash on delivery. The payment status updates itself when the gateway confirms.
5. Confirmation. A message with the order number and summary goes out automatically.
6. Status updates. Packed, handed to the courier, out for delivery, delivered: each change triggers a message, so nobody has to ask.

The key requirement is that the chat tool and the order system share one stock and one customer record. A tool that only sends messages is not enough; look for software where orders, inventory and payments live together, such as the inventory and order modules in StoreConsole, with the chat channel plugged in as another way to place an order.

WhatsApp Business Platform: the rules in August 2026

WhatsApp is the main chat channel for selling in most of the world, and its rules decide what you can automate. The ordinary WhatsApp Business app is fine for a very small shop. Once you want automation, catalogue messages and several agents on one number, you use the WhatsApp Business Platform through Meta or an approved provider. These are the rules that matter, as published in Meta's developer documentation.

Opt-in

You need a customer's consent before you send them business-initiated messages. Consent should be explicit, tell the person what kind of messages to expect, and be recorded. A checkbox at checkout, a click-to-chat button or a customer starting the conversation themselves are all common ways to collect it. Keep proof, and honour "stop" immediately.

The customer service window

When a customer messages you, a 24-hour customer service window opens. Inside it you can reply with free-form messages: text, images, product cards, anything. Every new message from the customer restarts the 24 hours. When the window closes, you can only reach that customer with an approved template message.

There is one more window: if the conversation starts from a Click-to-WhatsApp ad or a call-to-action button on your Facebook Page, a 72-hour free entry point window opens, and messages in it are not charged.

Template categories and pricing

Templates are pre-approved messages and come in three categories. Since 1 July 2025, Meta bills per delivered template message instead of per 24-hour conversation.

Message type	Typical use	Charged?
Free-form message inside the service window	Answering questions, sending product cards	No
Utility template inside the service window	Order confirmation right after a chat	No
Utility template outside the window	Shipping update a day later	Yes, per delivered message
Authentication template	One-time passcodes	Yes, per delivered message
Marketing template	Offers, restock notices, abandoned carts	Yes, always, at the highest rate
Any message inside a free entry point window	Replies to ad clicks within 72 hours	No

Rates depend on the customer's country and Meta revises them on a published schedule, so check the official pricing page before you budget. Two design lessons follow from the table. First, answer inside the window whenever you can, because it is free. Second, label templates honestly: a promotional message disguised as a utility template risks rejection or reclassification, and marketing is the category you pay most for.

A note on AI

Meta's Business Solution Terms were updated with effect from 15 January 2026 to stop AI providers from distributing general-purpose assistants through WhatsApp. Businesses that use AI to serve their own customers, for example to answer product and order questions, are not affected.

Instagram and Messenger follow similar logic

Instagram and Facebook Messenger work on the same principle: a customer-initiated conversation gives you a limited window, 24 hours as standard, to reply freely, and outside it your options are narrow. Because these chats often begin from a post, a story or an ad, the first message is usually "price?" with no context, so attach the product reference to the conversation as early as possible.

Whatever the channel, treat the platform as a place where you can be switched off. Keep your customer list, order history and consent records in your own system, not only in the chat app.

Payment: link or cash on delivery

The screenshot habit ends when payment is attached to the order. Two methods cover most cases.

• Payment link. The order generates a link for the exact amount. When the gateway confirms, the order moves to "paid" with no human involved. This removes fraud and matching work in one step.
• Cash on delivery. Still the default in many markets. Treat it as a status, not an afterthought: the order is marked COD, the courier collects, and the settlement is matched to the order later. Confirming the order in chat before dispatch (a quick "reply YES to confirm") cuts refusals at the door.

For high-value or first-time COD orders, a small advance payment by link is a reasonable filter.

Where AI helps, and where a person takes over

Most chat questions are repetitive. An AI assistant connected to your catalogue and orders can answer them at any hour, which matters when your customers write at night. The rule is to automate what is factual and low-risk, and hand off what involves judgement, money or emotion.

Situation	Who handles it	Why
Opening hours, delivery areas, return policy	AI	The answer is written down and does not change per customer
Is this in stock? What does it cost?	AI, reading live data	Must come from the system, never from the model's guess
Where is my order?	AI, reading order status	The most common question, and a pure lookup
Change the address, apply a discount	AI drafts, a person approves	Changes money or logistics
Complaint, damaged item, refund	Person	Needs empathy and authority
Large or custom order	Person	High value, negotiation

Two safeguards make this work. The assistant must answer stock, price and status only from live data, and say "I'll check with the team" when it does not know. And the handoff must carry the whole conversation to the person, so the customer never repeats themselves. Offer a human option clearly, in the first menu or the first message.

The numbers to track

Once chats become orders in a system, you can finally measure them. Four metrics are enough to start:

• Chat-to-order conversion: orders divided by new conversations, per channel and per source (ad, post, website button).
• First response time: median time from the customer's first message to the first reply. Compare conversion for chats answered within five minutes with slower ones.
• Average order value: compare chat orders with website orders. Product cards and suggested add-ons can lift it; check whether they do.
• "Where is my order" rate: share of orders that trigger a status question. If it is high, your automatic updates are not working.

An illustrative example: a shop receives 400 new conversations in a month and records 72 orders. That is an 18% conversion rate. If a change in response time lifts it to 21%, the same traffic produces 84 orders, with no extra ad spend. The point is not these figures but that you can now see the effect of every change.

Mistakes to avoid

• Messaging people who never opted in. It risks account quality ratings, blocks and, in some countries, legal trouble.
• Letting chat apps hold the only copy of your customers. Export and store contacts and consent in your own system.
• Selling without a stock reservation. The last unit will be promised twice.
• Using marketing templates for everything. You pay the highest rate and train customers to ignore you.
• An AI assistant with no way out. Always offer a human, and hand over the full history.
• Measuring nothing. Without conversion and response time, you will never know whether chat is profitable.

Back to the grey lamp at 11:40 pm. With the flow in place, the 9:10 question is answered from live stock, which shows zero, so the customer receives a back-in-stock alert instead of a payment request. Nobody compares a screenshot with a statement, because the payment link carries exactly 54.00. The owner opens the thread and finds an order number and a status, and nothing to apologise for.

Key takeaways

• Chat is a real sales channel, but a thread is not an order system: it has no stock, payment status or order number.
• Keep the conversation in chat and move order, stock and payment into software that shares one inventory and one customer record.
• On WhatsApp, get opt-in, reply inside the 24-hour window where it is free, and use utility templates for transactional updates; marketing templates are always charged.
• Replace payment screenshots with payment links, and treat cash on delivery as a tracked status with a confirmation step.
• Automate factual questions with live data, require approval for changes, and hand complaints and refunds to a person with the full history.
• Track chat-to-order conversion, first response time, average order value and the "where is my order" rate.

Anichur Rahaman is a software architect and the creator of StoreConsole. He designs commerce and ERP systems for growing businesses, with a focus on event-driven architecture, data integrity and self-hosted operations.