Skip to content
Chatbotscape
Editorial flat-vector illustration for Conversational Commerce Playbook: Which Funnel Step to Move Into the Thread First (2026)
14 min read

Conversational Commerce Playbook

Which Funnel Step to Move Into the Thread First (2026)

Quick answer: Conversational commerce is not one project. It is four separate moves, and almost every store that stalls does so because it attempted them in the order the software demo suggested rather than the order that actually works. This page is about sequence. Move order updates into the thread first, because that is where customers volunteer permission and the content costs nothing. Move pre-purchase questions second, once you have material worth answering with. Move reminders and recovery third, because they spend the permission and the content the first two moves produced. Move the cart and the payment last, and only if your market already buys that way. The glossary entry covers what conversational commerce is and which funnel steps chat can own; this page covers the order of operations and the test that tells you a move is finished before you start the next one.

The sequencing mistake, and why the demo encourages it

Watch any chat-commerce demo and it opens at the end: a catalog carousel, a product tapped, a payment confirmed, all inside a thread. It is the most impressive thirty seconds in the category, and it is the last thing you should build.

The reason is unglamorous. Every step of a chat funnel depends on two assets: an audience that has agreed to receive messages on that channel, and content good enough to answer with. A cart-in-thread flow assumes both. If you build it first, you have an elegant checkout serving nobody, because nobody opted in yet, and it answers no questions, because you never wrote the answers. The moves below are ordered by how much of each asset they consume versus produce. The first move produces permission. The second produces content. The third and fourth spend them.

Move one: order updates, because they buy the permission

Shipping notifications, delivery windows, and a "reply here if something's wrong" line are the least exciting thing on this page and the reason the rest of it becomes possible.

Three properties make them the correct opening. Customers actively want them, which makes the checkout opt-in the easiest permission you will ever ask for: you are offering information they already need rather than access to sell to them, and that opted-in list is the asset every later move draws on. The content cost is close to zero, because the messages are factual and templated rather than persuasive. And failure is cheap: a notification that arrives late is an annoyance, while a wrong answer about product compatibility is a return.

On WhatsApp this maps cleanly onto how the channel already works. Order updates fit the utility side of the template system, which is the category our broadcast campaign guide finds clears review faster and is scrutinized less strictly than marketing, and the customer's reply reopens a free-form window in which a person or a bot can handle the exception without another template. Template mechanics themselves belong to our broadcast campaign guide; what matters here is that the easiest messages to get approved are also the ones you should send first.

Build the exception path at the same time. Every order-update thread will receive replies that are not "thanks," and a status message with no human handoff behind it converts goodwill into a complaint. Our escalation playbook is the design reference.

What tells you this move is done: opt-in rate at checkout has settled at a number you can plan with, and the volume of "where is my order" contacts arriving through your other channels has visibly dropped.

Move two: the questions that stall a cart

The second move is pre-purchase question answering: sizing, materials, compatibility, stock, delivery time to a given address, returns policy. This is where chat has its clearest effect on conversion rate, and it is also the move most stores are least prepared for.

The dependency is content. A bot answering product questions is only as good as the material it retrieves from, which makes this move a knowledge base project with a chat interface attached. Our knowledge-base build guide is the prerequisite work, and the honest sequencing rule is that if your product information lives in three people's heads and one outdated PDF, this move is not blocked on software.

Wrong answers here are expensive in a way order updates never are, which is why this is the move that needs guardrails rather than enthusiasm. Ground the bot in your own product content instead of the model's general knowledge. Set a confidence line below which it hands off rather than guesses. And wire stock and price answers to the live source, not to a description someone pasted last quarter.

What tells you this move is done: the bot resolves product questions without escalation at a rate you would accept from a new hire, and your weekly transcript review stops surfacing confident wrong answers. Be careful with the conversion comparison: shoppers who ask questions are higher-intent before the bot says anything, so "sessions with a chat convert better" flatters the tool. The comparison worth running is against your own pre-launch baseline for the same product pages.

Move three: reminders and recovery

Only now does abandoned cart recovery make sense, and the reason is structural rather than stylistic. A recovery message sent to a list you did not earn is a broadcast with a cart attached, and a recovery message that cannot answer the objection behind the abandonment is a discount with extra steps. Move one supplies the permission. Move two supplies the answers the recovery thread needs when the customer replies "does it fit a 2019 model."

The sequence design, timing, incentive ladder, and measurement live in our abandoned cart recovery playbook, and the broader architecture of multi-message sequences lives in the drip campaign guide. What this page contributes is the placement: third, not first, and never before the exit conditions are written.

What tells you this move is done: recovery is producing orders at an incentive cost you would defend out loud, and opt-out rate on the channel has not moved.

Move four: cart and payment in the thread, and the market gate

This is the move the category is named after and the one most stores should skip.

Two conditions gate it. The first is market: your customers must already be people who transact in chat. That holds in commerce-first markets (India, Brazil, and Southeast Asia, per the pattern our conversational commerce entry sets out) and largely does not hold in the United States or most of Western Europe, where the thread convinces and the website closes. We are describing a consumer-habit pattern our reviews keep encountering, not a measured market share, so treat it as a prompt to check your own order data by country rather than as a statistic. The second condition is rails: your platform and channel must support the flow natively, because assembling a payment path out of links and manual confirmations reintroduces every friction the move was supposed to remove.

The glossary entry maps who does what across the funnel; the question here is which move each platform is worth buying for, and the answers our reviews support are narrower than the category's marketing.

AiSensy is the clearest move-four platform in our catalog, and it is India-specific. Our review verified catalog, single-product, and multi-product messages as native flow-builder blocks with UPI, card, and netbanking collection through Razorpay and PayU. It found no native Stripe or PayPal path, so an international seller gets the catalog blocks and routes payment through Shopify checkout or a custom gateway instead. Wati covers moves one through three on WhatsApp, with Shopify available as a separate paid add-on that our review priced at $4.99 per month at verification time; its Shopify commerce flow ends in a checkout deep-link rather than an in-conversation payment, which is the pattern that makes the general point. SendPulse is the exception to the split, with a caveat about how firmly we know it. The vendor documents native payment acceptance inside chatbot flows on its paid tier, which our review recorded from the pricing page rather than exercised, because that session ran on the free tier and the commerce scenario is still queued. What the review did confirm hands-on is the Stripe tooling in the integrations panel, with payment-link and invoice actions live on 27 May 2026. So the transaction step plausibly exists without a separate commerce platform behind it, it is not WhatsApp-catalog-shaped the way AiSensy's version is, and a buyer betting on it should exercise that path during the trial rather than take our word or the vendor's. Manychat is the strongest fit for moves one and three on Instagram and Messenger, and Tidio for move two on a widget-first store with native Shopify wiring. Each review records what was verified at its verification date, and those reviews carry affiliate links; our methodology sets out what that does and does not change about a score.

If you are outside a commerce-first market, the defensible version of move four is narrow: keep the cart in chat for reorders of things the customer has already bought, where there is nothing left to decide, and send everything else to the website.

What tells you this move is done: in-thread orders are a share of revenue you would name out loud without rounding up, and the payment path has run a full purchasing cycle without manual reconciliation. If neither is true after that cycle, remove it. An unused payment rail is a maintenance liability rather than an option you are holding.

The advance test

Do not start the next move until all four of these hold for the current one.

  • The number that move was supposed to change has changed, measured against a pre-launch baseline rather than a vendor dashboard that starts on day one of usage.
  • Support load did not increase. A move that shifts work from one queue to another has not paid for itself.
  • Opt-out rate on the channel is flat. Rising opt-outs mean you are spending the permission asset faster than you are building it, and every later move gets more expensive.
  • A named person owns the content that move depends on, with time booked to maintain it.

Run each move for at least a full purchasing cycle before judging it. Two weeks of an ecommerce funnel measures the week, not the change.

Which sequence fits which store

  • Physical goods, no chat presence yet: move one only, for a quarter. Collect opt-ins, ship notifications, and resist the roadmap until the list exists.
  • High-consideration or configurable products with heavy pre-sale questions: move two can legitimately go first, but only if the product content already exists in a form a bot can retrieve. If it does not, you are back to move one while you write it.
  • Existing opted-in list and a working support bot: move three is your highest-return next step, and our ecommerce chatbot shortlist is where to check platform fit for it.
  • WhatsApp-native audience in India: move four is real rather than aspirational, and AiSensy is the clearest fit in our catalog. In Brazil or Southeast Asia: treat it as conditional. None of the platforms above ships a verified native Pix rail inside the thread. On Wati and Manychat our reviews found Pix routing through the ecommerce platform rather than the chat vendor; SendPulse's payment node would reach it through a connected gateway, which its review flags as unverified per country. Ask a vendor which of those two shapes you would be buying, and verify the rail before you design the flow, because this is the capability that varies most between vendors and the one most likely to be oversold in a demo.
  • Low basket value, high volume, simple product: stop after move one plus basic FAQ deflection, and run the check rather than taking our word for it. Multiply your monthly unanswered or abandoned sessions by margin per order, and set that against the platform fee plus the hours someone spends keeping product content current. At low basket values that product is usually smaller than the maintenance line, which is what makes move two optional here and move four almost always wrong.
  • Services or appointments rather than goods: this is a different funnel; the lead generation chatbot pattern and our channel selection guide are the better starting points.

Frequently asked questions

Where do I start with conversational commerce if I have nothing set up?

Order notifications with a checkout opt-in, on whichever channel your customers already use. It is the cheapest thing to build, the easiest to get message approval for, and it produces the opted-in audience that every subsequent move depends on. Stores that start with a promotional flow instead spend months broadcasting to a list of a few hundred people who never agreed to hear from them.

Should I accept payments inside WhatsApp?

Not until moves one through three are done, and then only if two things are true: your customers already buy that way, and your platform has the rail natively on that channel. The documented pattern in our own reviews is the Wati one, where a Shopify commerce flow carries the customer to a checkout deep-link rather than through payment; SendPulse is the possible counter-example, since its vendor documentation puts payment acceptance inside the flow on the paid tier, though our review recorded that from the pricing page rather than exercising it, and it is not a WhatsApp catalog experience in any case. We have not measured what rebuilding a deep-link handoff as an in-thread flow returns, and we would not spend on a second payment rail without that number in hand.

Does conversational commerce only work on WhatsApp?

No, but the channel changes which moves are available. WhatsApp has the deepest commerce features and the strictest messaging rules; Instagram and Messenger are strong on discovery and recovery and weaker on transaction; a website widget can do question answering and cart assistance without a messaging opt-in. It is not consent-free, though. In most EU implementations the widget's own cookies fall inside your consent banner under the ePrivacy rules, and re-engaging that visitor later still needs a permission you have to collect. Check your own configuration with counsel rather than assuming the vendor handled it. Our channel selection guide walks the trade-offs, and the sequence in this playbook holds on any of them.

How long should each move run before I add the next?

At least one full purchasing cycle, which for most stores means a month rather than a fortnight, and longer if your product is seasonal. The advance test above is the real gate: the target number moved, support load flat, opt-outs flat, content owned. Stacking moves faster than that produces a funnel where nobody can say which part is working, and the usual outcome is that all of it gets blamed when one part is broken.

Which platform should I use for a chat commerce funnel?

The one that covers the moves you have actually reached, not the whole roadmap. For moves one through three, most mainstream platforms are adequate and the differentiator is your store integration and your content. For move four, the field narrows sharply to platforms built for commerce-first markets, and that is the point at which vendor choice genuinely matters. Our ecommerce chatbot shortlist and the Shopify integration guide cover the store-wiring side of the decision.

About this guide

Chatbotscape launched in 2026 as an independent review site for chatbot platforms. This guide is part of our SMB chatbot Academy. Its central recommendation, that most stores should stop after move one or two and skip in-thread payment entirely, argues against the feature set vendors most want to sell, and against our own interest in readers reaching for a platform upgrade. We have a mild commercial interest in readers choosing platforms through our reviews. To flag an error, write to editorial@chatbotscape.com.

Methodology

This page contains no new measurements. The WhatsApp 24-hour window and template-category mechanics trace upstream to Twilio's WhatsApp messaging documentation, fetch-verified on 23 July 2026 for our abandoned cart recovery entry and carried through the conversational commerce glossary entry, which also holds the provenance of the term itself in Chris Messina's January 2015 essay. Platform capability claims trace to those products' reviews and are stated as of their verification dates, 24 to 29 May 2026, rather than re-checked here. Most rest on hands-on sessions: AiSensy's catalog, single-product, and multi-product flow blocks with their Razorpay and PayU integration tiles, Wati's Shopify add-on and checkout deep-link, SendPulse's Stripe integration panel, and Manychat's and Tidio's channel behavior. One does not. SendPulse's in-flow payment acceptance on the paid tier is documented on the vendor's pricing page and was not exercised, because that review's session ran on the free tier, and this page flags it as such where it appears. The sequencing method, the advance test, and the store-type recommendations are editorial guidance rather than measured outcomes, and they follow our methodology. No conversion-rate or revenue-lift figures are quoted, because the ones circulating in this category are vendor-supplied and unreproducible.

Sources

  • Twilio Docs. Send WhatsApp notification messages with templates. twilio.com (24-hour customer service window and template mechanics; fetched and verified 23 July 2026 for our abandoned cart recovery entry, inherited here).
  • Messina, Chris. Conversational commerce. Medium, published 16 January 2015. medium.com (fetched and verified 25 July 2026 for our conversational commerce entry, inherited here).
  • Chatbotscape platform reviews for AiSensy, Wati, SendPulse, Manychat, and Tidio (hands-on verification 24 to 29 May 2026).
  • Chatbotscape evaluation methodology. /methodology (continuously updated).

Last updated

27 July 2026 — Initial publication aligned to methodology v3.12.1. Next scheduled refresh: 27 October 2026.