WhatsApp lead qualification works best as about four questions the lead answers with a tap, a simple score, and a person on every hot lead within minutes. It is the short structured chat that decides what happens to an inbound lead: who gets a call now, who books a slot, and who just gets a good answer. This guide is for founders and sales leads in Southeast Asia who run click-to-WhatsApp ads, or plan to, and want more of those chats to turn into sales.
Picture an illustrative case. A buyer taps your ad at 9:40 pm and types "price?". If the reply is seven paragraphs, or nothing until morning, that chat is over. If the reply answers in one line ("Plans start from [your published starting price] a month. Two quick taps and I can tell you which one fits."), then asks a question they answer with a thumb, you know by 9:42 whether a person should call tomorrow. Answer first, then ask. And make every question a tap, because a typed answer is where chats die. People type anyway, so keep a typed-text fallback and hand the chat to a person when the bot fails to understand twice.
At a glance: the numbers behind a qualification flow
Meta's limits are from Meta for Developers; the 24-hour window is the customer service window.
Why qualify leads inside WhatsApp instead of a web form?
Because the lead is already there, and the first minutes decide whether you reach them at all. A form sends the lead away and back again. A chat keeps them in the app they opened, and lets you ask, score and route before their attention moves on.
People say they prefer it. In a Meta-commissioned survey of 11,056 online adults in 22 markets, including Singapore, Indonesia, Malaysia, the Philippines, Thailand and Vietnam, 73.3% said they prefer messaging when they contact a business and 72.4% said they are more likely to buy from a brand that offers it. That is Meta's own research, not an independent audit.
Interactive chart
About 7 in 10 surveyed consumers favour messaging
Share of 11,056 online adults aged 18 to 64 in 22 markets who agree, between April and September 2025. Meta-commissioned research.
In Meta's commissioned survey, 73.3% prefer messaging when they contact a business, 72.4% are more likely to buy from a brand that offers it, and 67.7% agree a response from an AI chatbot is helpful.
| Statement | Agree |
|---|---|
| Prefer messaging to contact a business | 73.3% |
| More likely to buy from a brand that offers messaging | 72.4% |
| Agree an AI chatbot response is helpful | 67.7% |
Speed still matters, and a flow answers in seconds. It covers the evenings and weekends a sales team cannot, and Meta's customer service window gives you 24 hours from the person's last message to reply with free-form text. We found no published, vendor-independent figure for how much a faster first reply lifts qualification on WhatsApp, so we give none.
So which four questions, and in what words?
What should you ask, and in what order?
Ask what changes the routing, in the order a person would: need first, then timeline, then fit, then permission to follow up. Asking about budget first reads like a toll gate. Asking what they need first reads like help.
- Need. A list: "What are you looking for?" with up to 10 options.
- Timeline. Three buttons: this week, this month, just looking.
- Fit. Buttons or a list for size, location or segment. In Southeast Asia, the market or city is often the most useful fit question.
- Consent. Name the business, the topic and the way out: "Can [Business] send you WhatsApp messages about [your quote]? Reply STOP any time." Two buttons: "Yes, follow up" or "Not now". Store the answer, the exact wording shown and a timestamp in your CRM's consent record, not only in the chat log, so you can still prove it if the chat is deleted or you change systems. Never make the answer a condition of getting help.
Ask about budget or authority only if the answer changes how you route the lead, and then as a band, never a typed number. This is the classic BANT shape (budget, authority, need, timeline) reordered for a chat. BANT is sales practice, not a research finding, so adapt the order to your buyers.
Copy this flow
An illustrative four-message flow, within Meta's limits
1. Hi, thanks for messaging [BUSINESS]. I am an automated assistant, and a person can take over any time. What are you looking for? List button: See options. Rows: Get a quote / Pricing / Book a demo / Help with an order / Something else.
2. When would you like this sorted? Buttons: This week / This month / Just looking.
3. Where are you based? Buttons: Singapore / Malaysia / Elsewhere.
4. Can [BUSINESS] send you WhatsApp messages about [TOPIC]? Reply STOP any time. Buttons: Yes, follow up / Not now.
Routing: Get a quote goes straight to a person. Help with an order skips the rest and goes to support. Pricing gets your published prices in one line, then question 2.
Illustrative wording, not legal advice. Every button label is within Meta's 20-character limit and every list row title within its 24-character limit. Nothing you type leaves your browser.
"Help with an order" matters more than it looks. An existing customer should never be qualified like a stranger, so that row skips the questions and goes straight to support. "Get a quote" skips them too: a quote binds the business, so a person takes it.
Buttons, lists, Flows or free text: which should each question use?
Use the lightest component that can carry the answer: buttons for choices, lists for menus, a Flow for forms, free text for the lead's own words. Each has hard limits, and breaking them forces a worse format.
| Component | Meta's limit | Best for | Watch out |
|---|---|---|---|
| Reply buttons | Up to 3 buttons, labels up to 20 characters | Timeline, yes or no, consent | A fourth option means a list |
| List message | Up to 10 rows in total across up to 10 sections; row titles up to 24 characters | What they need, which product, which branch | Row descriptions up to 72 characters |
| WhatsApp Flow | Meta advises tasks under 5 minutes, one task per screen | 4 or more fields at once: a quote, an application | Sent outside a chat, it must sit in an approved template |
| Free text | None | A name, or the one question only they can phrase | Slowest to answer, hardest to score |
Sources: Meta for Developers on reply buttons, list messages, Flows best practices and sending a Flow.
Meta positions Flows for exactly this job: its documentation lists lead generation among their uses. Its best-practice guide adds two habits worth copying into any flow. Show progress, as in "Question 1 of 3", and end on a summary screen, so the lead sees what they told you before you act on it.
How much does a typed answer cost? Nobody publishes a clean drop-off rate per question type, so the builder below starts from illustrative rates you should replace with your own funnel numbers. Try one thing first: set typed answers as easy as a Flow screen, 90%. Four typed questions still keep only 656 of 1,000 chats, because every question compounds the loss. The number of questions hurts as much as the format.
Flow builder
How many chats does your question format keep?
Pick a flow, then change the rates to yours. How it is calculated is under the results.
How it is calculated: completed chats = chats x (button rate ^ button questions) x (list rate ^ list questions) x (Flow-screen rate ^ Flow screens) x (typed rate ^ typed questions), rounded to whole chats. Hot, warm and cold split the completed chats by the shares you set. Limit breaches count a button question with more than 3 options, a list with more than 10 rows and a Flow of more than 5 screens.
Illustrative defaults, not measured data: no published source gives completion rates per question type, so replace every rate with your own funnel numbers. The limits (3 buttons, 10 list rows) are Meta's. The 5-screen Flow flag is our rule of thumb for Meta's advice to keep a Flow under 5 minutes. Nothing you enter leaves your browser.
The same 1,000 chats, run through four flows at the calculator's illustrative rates, show how the format and the number of questions add up:
Interactive chart
At our assumed rates, taps keep over twice as many chats
Completed chats out of 1,000, by flow format, at the calculator's illustrative completion rates.
Illustrative: one list and three button questions keep about 789 completed chats per thousand started; a 4-screen Flow keeps about 656; adding three typed questions to the taps keeps about 333; four typed questions keep about 316.
| Flow format | Completed chats |
|---|---|
| 1 list and 3 button questions | 789 |
| A Flow of 4 screens | 656 |
| Taps plus 3 typed questions | 333 |
| 4 typed questions | 316 |
How do you score a lead and decide when a person takes over?
Give each answer a few points, route by the total, and let four triggers override the score at any time. The score is a sorting tool, not a verdict. Its job is to make sure the hot leads reach a person while they are still paying attention.
Four questions, three routes, one human rule
An illustrative qualification flow for inbound WhatsApp leads.
A simple illustrative scheme: zero to three points per answer, with "this week" and a clear need scoring highest. Hot leads go to a person now. Warm leads get a slot the bot offers. Cold leads get a useful answer and no follow-up unless they said yes to it. Calibrate thresholds against what closed last quarter.
The override triggers matter more than the points. Hand the chat to a person at once when the lead asks for a quote, a discount or any commitment (the bot can still state your published prices); asks a regulated question, such as financial or insurance advice; asks for a person; or when the bot has failed to understand them twice. That is the same rule we set out for designing the confirm step: an agent can gather and draft, but a person owns anything that binds the business.
Who owns the chat: bot until a trigger fires, then a person
An illustrative handoff rule for inbound WhatsApp leads.
After hours, "at once" means the bot says so plainly ("A person will reply by 9am tomorrow"), books the callback, and flags the chat hot so it is the first one answered.
One more rule keeps the bot on the right side of Meta. Section 4.7 (AI Providers) of Meta's Terms for the WhatsApp Business Platform bars general-purpose AI assistants where that is "the primary (rather than incidental or ancillary) functionality". In our reading, a bot that qualifies leads for your own business is incidental. That is not a safe harbour: the terms leave what counts to Meta's discretion. Our post on WhatsApp AI rules and real cost covers where that line sits, and the AI disclosure guide explains why you should say it is automated by the first reply.
What happens after the chat: follow-up, consent and telling Meta?
After the chat closes you need an approved template, the lead's consent, and a way to tell Meta which chats qualified. Three platform rules shape the follow-up, as Meta's pages read on 27 October 2026.
- The 24-hour window. When a user messages you, Meta opens a customer service window of 24 hours, reset by each new message from them. After it closes, "you can only send pre-approved template messages." Meta's non-template pricing page adds that from 1 October 2026 "Meta will charge on a per-message basis for all service messages". Meta's pricing page adds a free tier of 1,000 service messages per number each month, charged from the 1,001st; the non-template page does not mention it, so check your account. Rates: our rules and cost post.
- The free window from ads. If the lead came from a click-to-WhatsApp ad or a Facebook Page call-to-action on Meta's mobile apps and you reply within 24 hours, Meta's pricing page says a free entry point window opens, and while it is open you can send any type of message at no charge. The page says: "The FEP window may remain open for up to 7 days." (read 27 October 2026).
- Messaging limits. Meta's messaging limits count the unique people you message outside a customer service window in a rolling 24 hours, starting at 250 for a new business portfolio and rising through 2,000, 10,000 and 100,000 to unlimited. Qualification replies inside the window do not use that limit; template follow-ups do.
Consent is the step to get right in the flow itself. A reply inside a chat the lead opened is service. A promotional template sent days later to a Singapore number is usually a different matter.
Unless an exception applies, that later message needs a Do Not Call register result under 21 days old (section 43 of the PDPA, in the PDPC's Do Not Call guidelines, "Duty to check the DNC Register") or, in the words of the same guidelines ("Obtaining clear and unambiguous consent"), "clear and unambiguous consent": the person must have been told clearly and specifically that such messages would be sent to their number, and agreed through some positive action. The guidelines' Eighth Schedule list excludes some messages, such as those sent to an organisation rather than an individual, and those about the subject of an ongoing relationship, so a B2B follow-up may fall outside the rule. Check which applies to you; our post on the PDPA, consent and DNC sets out the test. This is general information, not legal advice.
Meta adds its own rule: its opt-in requirements say "Businesses are required to obtain opt-in before messaging people on WhatsApp", and set out what the opt-in must say:
Businesses must clearly state the business's name that a person is opting in to receive messages from.
Meta for Developers, "Getting opt-in", WhatsApp Business Platform documentation, as read 27 October 2026.
Then tell the ad which chats were worth it. Meta's Conversions API for Business Messaging accepts events such as LeadSubmitted and QualifiedLead from WhatsApp chats. The link back to the ad is the click ID, which Meta describes as "a personal identifier, unique per click", taken from the referral object in the message webhook. A common setup is to send LeadSubmitted when the flow completes and QualifiedLead when a lead scores hot or warm, so your ad delivery can use leads that qualified rather than every chat that started. Meta's page lists the events but does not promise a result, so test it. Treat the click ID as personal data and keep it only as long as your ad reporting needs it. Sharing conversion events with Meta for ad optimisation is a disclosure of personal data, so PDPC guidance says to notify leads of that purpose on or before disclosure, and consent may be needed too: say so in your privacy notice or first message, and check with counsel.
Interactive check
Is your WhatsApp qualification ready?
Tick what is already true. A guide, not an audit. Nothing you enter leaves your browser.
0of 8 basics in place
Close these gaps first
Bands: 0 to 3 not ready; 4 to 6 tighten the flow; 7 to 8 ready to scale. Based on Meta's platform rules and the practices in this post.
Where Origin Pi stands
We think WhatsApp is a strong place to qualify an inbound lead in Southeast Asia, and a poor place to leave one waiting. Our sales automation work builds that flow with a confirm step before anything that commits the business, and an audit trail of who was asked what and what they agreed to. For outbound, the companion piece is do AI SDRs actually work: inbound qualification is a quicker place to start, because the lead has already raised a hand.
Your first step: count how many WhatsApp chats you started last month and how many reached a person within ten minutes. If you cannot answer the second number, fix that before you change a single question.
Next step
Turn WhatsApp chats into qualified leads
Origin Pi
A governed WhatsApp qualification flow
We design the questions, scoring, handoff rules and consent capture, and connect qualified leads to your CRM and to Meta's Conversions API.
Ask for a qualification flow reviewleapbuzz
Click-to-WhatsApp and Meta ads
Fill the chat with the right leads through Origin Pi's marketing consultancy brand leapbuzz, which runs Meta advertising campaigns.
See leapbuzz Meta advertising services



