Types of conversation reopening on WhatsApp: reference table and what to do in each case
Why distinguish the type of reopening before responding
In a WhatsApp inbox, not all conversations that reopen are the same event. There are three situations that look alike from the outside —a chat that becomes active again— but that are governed by different rules: the customer who writes again on their own, the agent who picks up a case that had been paused, and the system that opens a new conversation with a template or a campaign. Confusing them is the most common cause of two opposite errors: sending a template when the customer had already written (and losing the thread), or replying outside the 24-hour window as if it were open (and getting an API rejection).
The table below organizes the three types and their variants. The 'trigger' column says who generated the activity; 'what it enables' says what can be done without a template; 'what to do' is the concrete action; and 'signal that it is handled correctly' is the closing check. There is no tool recommendation column: the table works the same with any provider that uses Meta's official API.
The 24-hour window is the one set by Meta: it opens with the customer's last incoming message and, while it is open, you can reply in free text. Outside that window, the only path is an approved template. Template approval times are up to 48 hours.
| Type | Trigger | What it enables | What to do | Signal that it is handled correctly |
|---|---|---|---|---|
| Customer returns within the window | The customer writes before 24 hours have passed since their last message | Reply in free text, without a template | Reply with the context of the previous case, not as if it were a new contact | The agent picks up the thread without asking again for the data the customer already gave |
| Customer returns outside the window | The customer writes after the 24-hour window has closed | A new window opens with their incoming message | Read the message, reply in free text within the new window and only afterwards, if needed, send a template | No template was sent before reading the incoming message |
| Agent picks up a paused case | No one wrote: the agent decides to return to a chat that was left open | Free text only if the window is still open; if not, template | Check first whether the window is open. If it is closed, choose an approved template and state the reason for returning | The customer understands why they are being written to and does not read it as spam |
| System reopens due to a campaign | A campaign or scheduled send reaches a contact who did not write | Approved template (HSM) only; free text does not apply | Use an approved template and leave the chat assigned to an agent before the customer replies | The conversation has an owner from the first minute, it does not remain orphaned |
| System reopens due to an auto-reply | An automatic rule replies outside business hours or by keyword | Free text within the current window | The auto-reply should say what will happen next (when a person will reply) and should not promise a timeframe that is not met | The customer does not write again asking whether they were attended to |
| Reopening via webhook or integration | An external flow (Make, Zapier, n8n, Slack) triggers a message or an action on the chat | Depends on whether the window is open at the time of the trigger | Check the time of the last incoming message before the flow sends anything; if it is closed, the flow must use a template | The flow does not attempt free text outside the window |
| Customer returns through another channel | The same customer writes through Instagram or Messenger instead of WhatsApp | Rules of the channel that receives, not of the one that was abandoned | Unify the contact's identity so as not to treat them as two different people | The history stays in a single contact, not duplicated |
| Reopening after audio transcription | The customer sends an audio and the system transcribes it | Reply in free text within the window | Use the transcription as context, but do not quote it verbatim if it may be mistranscribed | The reply does not repeat a phrase the customer did not say |
| Reopening of a chat already closed by the agent | The agent marked the case as resolved and the customer writes again | Free text within the window | Treat it as a continuation of the same case, not as a new inquiry, unless the topic is different | The data that is already in the history is not asked for again |
| Reopening with no reply from the customer | The agent or the system wrote and the customer did not reply | Nothing new: the window that was already open continues | Define a limit of attempts and do not keep insisting; if you need to return, use a template | The chat does not accumulate messages without a reply from the same side |
How to read the table without getting the column wrong
The first decision is always the same: who generated the activity? If the customer generated it, the window opens or renews itself and the agent's job is to reply with context. If the agent or the system generated it, the window does not open on its own: you have to check whether it is still current and, if not, go to a template. The 'what it enables' column is the one most often skipped and the one that generates the most problems, because many people assume that having the chat open on screen is the same as having the window open in the API. It is not the same.
The second decision is whether the case is a continuation or a new inquiry. A reopening about the same topic is handled with the history in view; a reopening about a different topic should be treated as a new case even if the contact is the same. This does not change the technical rule of the window, but it changes how the chat is assigned and how it is measured afterwards.
- If the customer wrote last: the window is open or just opened. Free text.
- If the last message is from the business and more than 24 hours have passed: window closed. Template.
- If the last message is from the business and less than 24 hours have passed: window open. Free text, but be careful not to sound pushy.
- If the activity was generated by a campaign: the conversation starts with a template and needs an assigned owner before the customer's first reply.
Common mistakes when reopening a conversation
The most costly mistake is sending a template to someone who just wrote. Besides being unnecessary, it breaks the thread: the customer sees a message formatted as a notice when they were expecting an answer to what they asked. The second mistake is the opposite: trying free text outside the window and hitting the API rejection, which is usually discovered late, when the customer has already written asking why no one replied.
- Sending a template when the customer already wrote: context is lost and a send that wasn't needed is wasted.
- Trying free text outside the window: the API rejects it and the message doesn't arrive.
- Reopening with a campaign without assigning the chat: the customer replies and no one is watching that conversation.
- Treating a reopening for a new topic as a continuation: context that doesn't apply gets carried over.
- Insisting without limit when the customer doesn't reply: turns a follow-up into perceived spam.
- Duplicating the contact because they came back through another channel: the history is lost and they're asked for data they already gave.
None of these mistakes is fixed with a general rule: they depend on whether the window is open and on who generated the activity. That's why it's best to resolve them in the order of the table, not by intuition.
What to check before reopening: a four-step checklist
- 1Look at who sent the last message and at what time. That defines whether the window is open.
- 2If you're the one generating the activity, check whether there's an approved template for that reason before writing anything.
- 3Assign an owner to the chat before the customer replies, especially if the reopening came from a campaign.
- 4Record the reason for the reopening in an internal note or tag, so the next agent doesn't have to guess it.
The fourth step is the one most often skipped and the one that saves the most time later. A reopening with no recorded reason forces the next agent to read the entire history to understand why the chat became active again. A tag or a short note solves that in seconds.
How to measure whether reopenings are being handled well
You don't need a complicated dashboard. It's enough to count, over a sample of reopened conversations, how many were resolved without sending an unnecessary template, how many had an assigned owner before the customer's first reply, and how many were closed without the customer asking again whether they were attended to. Those three numbers are enough to know whether the table's criteria are being applied or whether each agent is handling it their own way.
The first three indicators can be pulled from the inbox and from message and campaign statistics. The fourth requires a reason tag when reopening.
| Indicator | How it's calculated | What to look at if it goes wrong |
|---|---|---|
| Unnecessary templates | Reopenings where a template was sent while the window was open, over total reopenings | If it's high, the team isn't checking who sent the last message |
| Orphan chats | Campaign reopenings without an assigned owner before the customer's first reply | If it's high, automatic assignment when the campaign is triggered is missing |
| Support re-inquiries | Customers who write again asking whether they were attended to, over total reopenings | If it's high, the auto-reply or the first message isn't saying what happens next |
| Reopenings for a new topic treated as a continuation | Cases where the reason for the reopening doesn't match the topic of the previous history | If it's high, there's no explicit criterion for when to open a new case |
How this relates to templates and campaigns
Templates (HSM) are the only way to reopen a conversation once the 24-hour window has closed. That means any reopening generated by the business —a follow-up, a campaign, a notice— depends on having an approved template for that reason. Approval takes up to 48 hours, so a campaign put together the same day can't depend on a template that isn't approved yet.
Meta's conversation prices vary by country and are not set by the inbox provider. If the cost of reopening with a template is a variable for your operation, it's worth checking Meta's official documentation for your country rather than a third-party table that may be outdated. The decision of when it's worth reopening with a template and when it's better to wait for the customer to write is, at bottom, a decision of cost versus urgency.
Where a tool like Wando comes in
This type of distinction is what a tray with AI suggestions helps sustain without each agent having to remember the rule: the suggestion is built with the context of the history and the agent approves it with one click, so the case of 'customer returns within the window' is not answered as if it were a new contact. Wando uses Meta's official WhatsApp API, learns from the business's real responses and is designed for LatAm SMEs. The free trial lasts 14 days and does not ask for a card; current plans and prices are at wando.online/pricing.
Edge cases worth having resolved in advance
There are situations that do not fit cleanly into any row of the table and that are worth deciding before they happen, not in the moment. For example: the customer writes exactly at the minute the window closes; the agent picks up a case that another agent had already closed; or the system triggers a campaign to a contact who wrote two hours ago. In all three cases the underlying rule is the same — look at the last inbound message — but the operational decision changes depending on who has the chat assigned.
- Customer writes at the window boundary: treat as open window and reply in free text; if the API rejects it, go to a template without arguing with the system.
- Two agents on the same reopened chat: define a single owner before replying so as not to send two different messages.
- Campaign triggered to a contact who wrote recently: check if the window is still open and, if so, assess whether the campaign is needed or if free text is enough.
- Customer who returns with a complaint after a campaign: prioritize the complaint over the reason for the campaign and do not mix the two threads in the same message.
None of these edge cases needs a new feature: they need the team to have written down what to do. The table above is the starting point; edge cases are the boundary where the table falls short and it is worth adding a row of your own.
Frequently asked questions
What is the 24-hour window and how does it affect a reopening?+
It's the period set by Meta: it opens with the customer's last inbound message and, while it's active, you can reply in free text. When it closes, the only way to reopen the conversation is an approved template. That's why the first step in any reopening is to check who sent the last message and at what time.
If the customer writes to me after ten days, can I reply without a template?+
Yes. The customer's inbound message opens a new window, so you can reply in free text within that window. What you can't do is send a template before reading what they wrote: that breaks the thread and wastes an unnecessary send.
When is it better to use a template to pick up a chat?+
When the movement is generated by the business and the 24-hour window is closed. There the template is the only way. If the window is still open, free text with context is better, even if the reason is a follow-up.
How long does Meta take to approve a template?+
Up to 48 hours. There is no more precise deadline that Meta guarantees, so a campaign that depends on a new template should not be scheduled for the same day.
What do I do if a campaign reopens a chat and nobody is watching it?+
Assign an owner before the trigger, not after the customer replies. A chat reopened by a campaign without an assigned agent is the most common cause of the customer writing to ask if they were attended to.
How do I tell a reopening about the same topic from a new inquiry?+
By looking at the reason for the inbound message against the history. If it's the same topic, it's resolved with the context in view; if it's different, it's better to treat it as a new case even if the contact is the same, especially for assignment and later measurement.
What happens if the customer comes back through Instagram or Messenger instead of WhatsApp?+
The rules of the channel that receives the message apply, not those of the one that was abandoned. The important thing is to unify the contact's identity so you don't end up with two separate histories for the same person.
How do I know if the team is handling reopenings well?+
With three numbers on a sample: how many reopenings were resolved without sending an unnecessary template, how many had an assigned owner before the customer's first reply, and how many were closed without the customer writing again to ask if they were attended to. If all three are low, the table's criteria are not being applied.
Answer WhatsApp with AI
Try Wando free. No credit card required.