What This Document Is
Most e-commerce brands treat the sale as the finish line. Pay for the ad, close the order, ship the jar, done. Then they pay full price for the next customer while the last one quietly forgets them.
This playbook runs the whole lifecycle, step by step:
- A source-aware first reply that captures every lead, including the retail buyers you never normally see.
- Consultative AI selling that diagnoses before it recommends.
- A close that happens inside the chat, payment, delivery details, everything.
- Fulfilment updates that send themselves.
- A reorder engine that reaches out before the product runs out, before the birthday, before the festive season.
- An AI Manager loop that studies every conversation and tells you what to change.
One example business is used throughout: a honey brand. The same flow works for any consumable or repeat-purchase product, supplements, coffee, skincare, pet food, corporate gifting. Copy the structure. Swap the product.
The Six Phases at a Glance
- Phase 1: Instant Capture and the Source-Aware First Reply
- Phase 2: Consultative Selling (Diagnose Before You Recommend)
- Phase 3: The In-Chat Close
- Phase 4: Fulfilment That Updates Itself
- Phase 5: The Reorder Engine (Where the Margin Lives)
- Phase 6: AI Manager Analysis and the Level-Up Loop
- Appendix: Full Execution Checklist
Phase 1: Instant Capture and the Source-Aware First Reply
Step 1.1: Reply in Seconds, at Any Hour
Someone scrolling Facebook at 11pm sees your ad, taps it, and lands in your WhatsApp. The conversation starts while the interest is still hot. Nobody on your team wakes up. No lead sits in an inbox until morning.
Speed is not a nice-to-have in e-commerce. The customer who messages three brands buys from the one that answers.
Step 1.2: Know Where the Lead Came From, and Greet Accordingly
Before the customer types anything, the AI already knows which ad and which campaign brought them in. The first message changes with the source:
- Ad lead: greeted warmly and asked what they are looking for, in the language of the ad they clicked.
- QR code on the jar: a customer who picked up your product at a supermarket scans the code and gets thanked for the purchase and asked how they are enjoying it.
- Referral or returning customer: recognised and welcomed back, never treated like a stranger.
The QR-on-packaging move deserves special attention. A retail buyer would normally stay invisible to you forever, the supermarket owns that relationship. One QR code on the label turns them into a customer you can talk to, follow up with, and resell to.
Generate the QR code inside the ABC Sales AI app itself, it opens WhatsApp with the message pre-typed, so scanning takes one tap. And give people a reason to scan: a usage guide, not a sales pitch.
Scan for the complete guide: how to store your raw honey, 5 ways to take it daily, and what makes Tualang different.
They scan to use the product better. You gain a conversation, a profile, and a reorder cycle. Print it on every jar, box, and bag you ship.
Step 1.3: Open with a Choice, Not a Blank Question
Do not open with "How can I help you?" Make answering effortless:
Hi [Name], welcome to [Brand]. What brings you here today?
A) Looking for a recommendation
B) Price and promo info
C) Reordering my usual
D) Question about my order
Every option routes into a different flow, and every answer is a data point in the customer profile that starts building from message one.
Phase 2: Consultative Selling (Diagnose Before You Recommend)
Step 2.1: Ask Before You Pitch
The customer says they keep getting sore throats and heard raw honey helps. The AI does not dump a product catalogue on them. It asks first:
- Is this for you or for the family?
- How do you usually take honey?
- Any kids at home?
Only after understanding the situation does it recommend: a robust Tualang for their immunity, a milder Acacia the kids will actually take. The customer feels advised, not sold to.
Step 2.2: Answer Hard Questions from the Knowledge Base
Certifications, origin, suitability for diabetics, halal status, shipping zones: the AI answers accurately from your knowledge base, always current because you update it in one place. No more "let me check with my boss" delays that kill momentum at the exact moment the customer is ready.
And when a picture answers better than a paragraph, the AI knows: ask about the texture and it sends the product video, ask whether it is really raw and it sends the lab certificate PDF, ask what the gift box looks like and the photo arrives. That is Smart Media Reply, the AI detects which photo, video, or PDF the moment calls for and sends exactly that one.
Step 2.3: Build the Profile While You Talk
While the conversation runs, the AI quietly accumulates the customer profile: their needs, their household, their preferences, how fast they will likely finish a jar. Nothing gets typed into a CRM by anyone. It builds itself, conversation by conversation, and every later phase of this playbook runs on it.
Phase 3: The In-Chat Close
Step 3.1: Tabulate and Quote Inside the Conversation
The customer decides: one jar Tualang, one jar Acacia. The AI tabulates the order, jars, promo, shipping, and quotes the total. No cart link that half of them never open. The checkout is the chat.
Step 3.2: Take Payment the Way Your Customer Pays
The AI shares your payment route: bank-in details, a Stripe link, or your Shopify checkout. When the customer sends back a payment screenshot, the AI verifies it and confirms the order. Then it collects the delivery details, and the order lands in your system.
From "I want to buy" to "order confirmed", no human touched anything.
Step 3.3: Tell Meta Who Actually Buys
The moment a purchase is confirmed, an AI tool fires the sale back to Meta through the Conversion API.
This closes the loop your ads have been missing: Meta normally only knows who clicked and messaged, so it optimises for people who inquire. Once every confirmed purchase flows back as a conversion event, Meta learns what a buyer looks like and starts sending you more of them, same budget, better leads.
Setup is a one-time job (see the Meta Conversion API setup guide); after that, every in-chat sale trains your ad account automatically.
Step 3.4: Recover the Ones Who Stall
If the customer goes quiet before paying, a follow-up goes out on schedule: a before-and-after testimonial from another customer, or a simple "hey, still keen on the honey?" The tone stays helpful, never desperate. A stalled checkout is not a lost sale, it is a lead in the follow-up automation.
Phase 4: Fulfilment That Updates Itself
Step 4.1: Orders Flow to the Warehouse Automatically
The confirmed order flows straight to fulfilment. When the courier assigns a tracking number, the system syncs it, and the AI messages the customer that their honey has shipped, tracking number included, with a small tip about storing it away from sunlight. Useful, human, on brand.
Step 4.2: Check In After Delivery
A few days after delivery, the AI asks: did everything arrive safely? If yes, wonderful, the relationship deepens. If something is wrong, a human gets pulled in immediately, before the customer writes the one-star review.
Phase 5: The Reorder Engine (Where the Margin Lives)
Step 5.1: Understand the Uncomfortable Truth
Acquiring that customer was expensive, and ads keep getting pricier. Most brands pay all that money, close one order, and let the relationship die. The customer finishes the jar, forgets the brand, and buys whatever is on the shelf next time.
The AI captured the purchase date. It never forgets. This is where it quietly becomes your most profitable salesperson.
Step 5.2: Reach Out Before the Product Runs Out
A jar of honey lasts about six weeks. Just before the customer runs out, the AI reaches out: running low? Want me to send your usual? Because it remembers their exact order, address, and payment preference, the reorder closes in a few messages.
Set the consumption cycle per product: six weeks for a jar of honey, four for coffee, eight for a supplement tub. The reminder lands right before the shelf goes empty.
Step 5.3: Work the Calendar: Birthdays and Repeat Occasions
- A month before the customer's birthday, the AI reaches out with a treat-yourself offer or a gift idea.
- If there is a known cycle, the corporate gift sets they ordered last Chinese New Year, the festive family order, any repeat occasion, the AI reaches out a month before it comes around again, before they have arranged it elsewhere.
New customers cost a fortune. These follow-ups cost nothing. This is where the margin lives.
Step 5.4: Returning Customers Never Start From Zero
The customer messages again after three months. They do not re-explain who they are, what they ordered, or where they live. The AI remembers their details, preferences, and history, so picking up their usual order takes a few messages instead of a whole new conversation.
Step 5.5: Capture the Referral in the Same Breath
They mention their mother wants to try it too, and she is diabetic. The AI recommends the suitable option, creates a new profile for her linked as a family referral, and now there are two reorder cycles running instead of one. Word-of-mouth growth, captured automatically instead of lost in a chat log.
Phase 6: AI Manager Analysis and the Level-Up Loop
Step 6.1: Schedule Your Morning Briefing
While all of this runs, the AI Manager is doing something no human team can: reading every single conversation. Schedule a report on the cadence you want, every morning if you like, covering new leads, sales, who needs follow-up, and whose jar runs low today.
And when you have a question, which ads bring actual buyers, what customers keep asking about, you just ask in plain language.
Step 6.2: Read the Level-Up Report and Act On It
On a regular cycle, the AI Manager hands you a Level-Up Report:
- What is working in your sales conversations.
- The objections that keep going unanswered.
- Where customers drop off.
- Which customer segments are most valuable.
- Specific recommendations on what to change.
You apply the changes, and the next report tells you whether they worked. That is the flywheel: sell, analyse, learn, adjust, sell better. A human team improves slowly and forgets when staff leave. This system compounds, and keeps every lesson forever.
Step 6.3: Mine the Conversations for Demand You Do Not Serve Yet
The chats also contain expansion signals. Customers keep asking for a product size you do not stock, a flavour you do not carry, or delivery to a state you do not cover. AI Manager counts the pattern, surfaces it, and suggests a tag, for example asked-gift-box, so the day you launch it, you announce to a list of people who asked for it themselves.
Step 6.4: When It Works, Clone It
Once the honey playbook is proven, launching your second brand does not start from zero. New AI personality, new product knowledge, same proven flow underneath. Days to launch, not months. And because customer profiles are unified, you know each customer's full value across every brand you run.
The Whole Framework in One Sentence
If you only remember one thing, remember this:
A lead arrives, the AI consults and closes, the order fulfils itself, the AI follows up before the jar runs out, before the birthday, before the festive season, the customer reorders, upgrades, and refers, and the AI Manager reports how to do it all better next month.
Every step runs 24/7 without adding headcount. The part of the business most brands throw away, the repeat purchase, becomes the engine.
When You Outgrow This Playbook
Everything in this playbook runs on the platform included in every plan, you can build it yourself. But growing brands tend to hit the same next walls, and those are the systems our team hand-installs end to end on the Scale plan:
- Order volume outgrows chat-by-chat management: the E-commerce Lifecycle install automates inquiry to fulfilment to reorder as one wired system, Shopify, payments, tracking, and all.
- Content demands never stop: the Content Engine researches, designs, writes, and posts your social content on autopilot, from your templates and tone.
- Ad spend grows faster than your attention: the AI Ads Manager watches your Meta ads 24/7, scales winners, kills losers, and assigns the next creative shoot.
- Running the brand across too many tools: Connected Business Pulse wires POS, store, sheets, and CRM into your team chat with AI reading it all.
If two or more of those sound familiar, book a strategy call and ask about the Catalogue.
Appendix: Full Execution Checklist
A. Capture Setup (Phase 1)
- AI replies within seconds on every connected channel, at any hour.
- Ad and campaign source is captured for every lead.
- First message adapts to source (ad lead, QR scan, referral, returning customer).
- QR code is generated in the ABC Sales AI app (message pre-typed, one tap to send) and printed on every jar, box, and bag.
- The QR offers a usage guide as the lead magnet (how to store it, how to use it daily), not a sales pitch.
- Opening message offers A/B/C/D choices instead of an open-ended question.
B. Consultative Selling (Phase 2)
- The AI asks diagnosis questions before recommending.
- Product recommendations map to stated needs (who it is for, how it is used).
- Knowledge base covers certifications, origin, suitability, halal, shipping zones, and is updated in one place.
- Customer profile builds automatically: needs, household, preferences, consumption pace.
C. In-Chat Close (Phase 3)
- The AI tabulates orders and quotes totals inside the chat.
- Payment route is wired: bank-in, Stripe link, or Shopify checkout.
- Payment screenshot verification and order confirmation are automatic.
- Delivery details are collected in-chat and the order lands in your system.
- Meta Conversion API tool fires every confirmed purchase back to Meta (one-time setup), so ads optimise for buyers, not just inquirers.
- Stalled-checkout follow-up is scheduled (testimonial or gentle nudge).
D. Fulfilment (Phase 4)
- Confirmed orders flow to the warehouse automatically.
- Tracking numbers sync and the shipped message fires with a care tip.
- Post-delivery check-in is scheduled.
- Problem replies escalate to a human immediately.
E. Reorder Engine (Phase 5)
- Consumption cycle is set per product (e.g., 6 weeks per jar).
- Running-low reminder fires before the product runs out.
- Birthday outreach is scheduled a month ahead.
- Repeat occasions (festive orders, corporate gifting cycles) are tagged and scheduled a month ahead.
- Returning customers are recognised, with history and preferences loaded.
- Referrals get their own linked profile and their own reorder cycle.
F. AI Manager Loop (Phase 6)
- Morning briefing is scheduled (leads, sales, follow-ups, running-low list).
- Level-Up Report cadence is set, and recommendations are reviewed and applied.
- Unmet-demand patterns are surfaced and tagged for future launches.
- Second-brand cloning checklist exists: new personality, new knowledge base, same flow.
G. Strategic Health Checks
- Every first reply matches the source that brought the lead in.
- The AI diagnoses before it recommends, always.
- No order requires a human touch between "I want to buy" and "order confirmed".
- Every purchase date, birthday, and occasion is captured, the reorder engine runs on them.
- The Level-Up Report is read and acted on, the flywheel only compounds if you turn it.
End of Playbook.