Send a WhatsApp template to many customers at once

Your bot answers people who message it. Broadcasts are the other direction: reaching out to many customers at once with an approved template — a seasonal promotion, a reminder that bookings open Monday, a notice that you’re closed next week. You’ll find Broadcasts as a tab at the top of your project, next to Conversations. Inside it are three sub-tabs: Broadcasts lists your campaigns and is where you compose new ones, Segments is where you save reusable audiences, and Contacts is your address book — everyone you can broadcast to, and where you tag people and manage your tags.

​

Before you start

You need two things:

1. An approved template. WhatsApp doesn’t let businesses send whatever they like to people who haven’t messaged recently — every broadcast goes out as a template that Meta has reviewed and approved. Approval takes minutes for a utility template and can take hours for a marketing one, so do this first.
2. Contacts to send to. Either from people who’ve messaged your bot, or imported in bulk.

​

Building a broadcast

Click New Broadcast and you’ll go through six steps.

​

Launching and scheduling

You send a broadcast one of two ways:

• Launch it now. A finished draft has a Launch action in the broadcast list — click it, confirm, and the broadcast goes out straight away.
• Schedule it. Give it a date and time on the Schedule step and Paige launches it for you when that time comes.

​

Sending in batches (waves)

If you have a large audience and want to test the waters first, switch on Send in batches (waves) on the wizard’s Schedule step and choose how many to send now. When the broadcast launches, Paige freezes the whole audience and sends only that first batch — everyone else is put on hold. Your counters still cover the full audience on the one broadcast, so you never have to stitch several broadcasts together. The broadcast then shows as Partially sent rather than jumping to Completed. Open it and a panel tells you how many are still held — 200 recipients remaining, say — with a Send now / Schedule toggle for releasing them. Release as many waves as you like, whenever you like. Each time, anyone who has opted out since you launched is skipped automatically. The broadcast only marks itself Completed once the last held recipient has been released and finished sending.

​

Releasing a wave now

Leave the toggle on Send now, then either:

• Type how many to release and click the send button, which counts back what you typed — Send 50 more.
• Or click Send all remaining to release everyone left.

​

Scheduling a wave for later

Switch the toggle to Schedule and you get two fields instead:

• How many — leave it blank to schedule everyone remaining (the placeholder reads All), or type a number.
• Send at (your local time) — the date and time to release it. It has to be in the future.

The submit button counts back what you typed — Schedule 50 — or reads Schedule all remaining when you left the count blank. Paige notes underneath that it fires within ~15 minutes of the chosen time: the release runs on a fifteen-minute cycle, so treat the time you pick as “no earlier than”, not to the second. Once scheduled, the controls are replaced by a banner — 50 scheduled for 14 Aug 2026, 09:00, or All remaining scheduled for that time — with the note that it releases automatically around this time (within ~15 minutes) and that opted-out contacts are skipped at release. To change a pending schedule, just schedule again: setting a new count or time overwrites the previous one. There’s no separate edit control. Scheduling a wave is only offered while a broadcast is Sending or Partially sent and still has recipients held back. It isn’t available on self-hosted projects. Batched sending is available on standard Paige-hosted projects. Projects running their own self-hosted backend send everyone at once, and the toggle is hidden.

​

What it costs

Paige doesn’t charge for broadcasts, and they don’t use your credits. Meta bills your connected number directly for the messages, and the estimate on the Confirm step is a rough guide rather than a quote. The one limit is a daily message cap on each project, which starts at 1,000 messages a day and resets at midnight UTC. It’s a per-project setting rather than a fixed platform ceiling, so a project can be raised above the default. There’s no self-service control for it — ask the Help Agent if 1,000 a day isn’t enough for you. You can see where you stand without doing the arithmetic: the Broadcasts tab always shows how much of today’s allowance you’ve used, and the wizard’s Confirm step tells you how much is left before you commit. If your audience is bigger than the remainder, you know up front that the overflow goes out the next day.

​

When a broadcast hits the cap

A broadcast that runs out of allowance mid-send pauses visibly and picks up where it left off. You get a notification, its badge changes from Sending to Paused — daily limit, and a banner on both the broadcast list and the broadcast itself says how many messages you’ve used and the exact time the rest will start going out. Nothing is lost — the held recipients are still queued, and they go out on their own when the allowance resets. There’s nothing to click.

​

Who actually receives it

Two groups are always excluded, whatever your filter says:

• Anyone who opted out. This is checked when the broadcast sends, not when you build it — so someone who sends STOP after you schedule is still excluded.
• Numbers confirmed not to be on WhatsApp.

That second one is worth understanding. Paige doesn’t know whether an imported number is on WhatsApp until it tries. Your first broadcast to fresh imports is what finds out: the dead ones fail once with Not on WhatsApp, and they’re quietly left out of every audience after that. A first broadcast with more failures than you expected usually means exactly this, and it’s self-correcting.

​

Watching it go out

Open a broadcast to see live counters updating as it goes. Total is the size of the whole audience, and next to it sit six status tiles: Queued, Sent, Delivered, Read, Failed, and Canceled. There’s no “Held” tile. If you sent in batches, the recipients still waiting are shown in the wave panel below the tiles — 200 recipients remaining — and that’s the number to look for, not one of the six. Below that, every recipient is listed with their status. Failures say why in plain language:

​

When a campaign reached nobody

If a broadcast finishes without matching a single contact, it’s marked Reached nobody — as a chip on the Broadcasts list and as a banner on its own dashboard. It’s called out explicitly because an empty send otherwise looks exactly like a successful one: a finished campaign, a green tick, no failures. The difference is that nothing happened. This almost always means the audience didn’t select anyone — a tag no contact actually carries, a name or number filter that matches nothing, or everyone who did match having opted out. Nothing was sent, so nothing was charged and none of your daily WhatsApp allowance was used. Open the segment, check its filters against your Contacts list, and send again once it selects the people you expect.

​

Downloading the recipient list

Next to the filter is a Download CSV button. It saves the recipient list as a spreadsheet you can open in Excel or Google Sheets, or feed into another system — each person’s name, number, WhatsApp status, failure reason if any, the wave they went out in, how many attempts were made, and the times things happened. It respects the filter you have on. With Failed selected you get only the failures, which is the quickest way to work through bad numbers offline. Phone numbers keep their leading + rather than being mangled into plain numbers by the spreadsheet. The file is a snapshot of that exact moment — deliveries and reads keep arriving from WhatsApp for a while after a send — so the filename carries the date and time, and you can download again later for an updated picture. Two cases where you won’t get a file:

• A broadcast that hasn’t started sending. There’s no recipient list yet to export.
• A broadcast that failed before it built its list. You get a short explanation instead. An empty spreadsheet would look like nobody was in your audience, when in fact the send never got started.

Self-hosted projects keep their recipient list on their own backend rather than in Paige, so the download isn’t offered there.

​

Exporting several broadcasts at once

For a monthly report, you don’t need to open each campaign. Export sits at the top of the Broadcasts list, next to New Broadcast. Tick boxes aren’t cluttering up the rows until you need them — click Export (or Select) and they appear. Tick the campaigns you want, then confirm. The button stays greyed out until you’ve picked at least one, and tells you so if you hover it. A window then asks what to include.

• Campaign summary — one spreadsheet row per campaign with its status, template, audience, every count (total, queued, sent, delivered, read, failed, canceled, held), the estimated cost, and all the timestamps.
• Recipient detail — a ZIP containing one recipient list per campaign (the same file Download CSV gives you), plus the summary if you ticked it, plus a short README.

You can tick up to 25 broadcasts at a time, and there’s a ceiling on the total recipients one export may cover. Go over either and the window tells you, so you untick a few. Campaigns that haven’t started sending still appear in the summary but contribute no recipient list, because there is nothing to list. The window names them before you download and the README names them again. Like the single download, bulk export is owner and workspace-admin only. On a self-hosted project the ZIP arrives with the summary and the README but no recipient files, and the README says so for each campaign — you still get a usable report, it just stops at the campaign level.

​

Cancelling and retrying

Cancel appears on a broadcast that is Scheduled or Sending, and stops it. Anything already sent has left and can’t be recalled — cancelling only stops what’s still queued, including anyone held back for a later wave. It can’t be undone. You’ll be asked to confirm, and the confirmation counts exactly who is affected — the queued recipients plus any held for a later wave. Since cancelling is final, and sending the same campaign again means building a new broadcast, it’s worth reading those numbers before you go ahead. Retry failures appears once a broadcast has finished with failures. It re-queues only the ones worth retrying — a rate-limited send, say — and skips the permanent ones like Not on WhatsApp or Template rejected, since retrying those would just fail again.

​

Reusing a broadcast

Sending much the same campaign again — a monthly newsletter, a weekly special — doesn’t mean rebuilding it. A finished broadcast has a Duplicate action that opens the wizard already filled in with what the original used. It isn’t a blind copy. Paige re-checks everything against today: a template that’s since been rejected, a segment that’s been deleted, a media file that’s gone. Anything that no longer holds up is flagged for you to fix rather than silently carried over — which is the failure mode a plain copy would have. Once you do save, the copy is an ordinary draft: it needs launching or scheduling like any other, and it counts against the same daily cap and passes the same checks. Clicking Duplicate twice by accident won’t give you two copies.

​

Auto-retargeting quality-limit failures

WhatsApp sometimes blocks part of a broadcast because of your number’s quality rating, or because a particular person has already received too many messages from you. Those recipients show as Blocked by quality limits, and they’re usually reachable again once a day has passed. Auto-retarget quality-limit failures on the wizard’s Schedule step does that for you. It’s off by default. The switch explains itself in-product:

If WhatsApp blocks some recipients for quality or per-user messaging limits, automatically try them again about 24 hours later. You can cancel the retry any time before it sends.

Here’s what actually happens when it’s on:

• About 24 hours after the original broadcast finishes, Paige creates a follow-up broadcast aimed at exactly the recipients WhatsApp quality-blocked.
• It uses the same template and the same variable settings as the original.
• The follow-up appears in your broadcast list under the original’s name with (auto-retarget) on the end — Winter sale (auto-retarget) — in Scheduled status. You can open it, check it, and Cancel it like any other scheduled broadcast before it sends.

Three limits worth knowing:

• It happens once. A retarget is never itself retargeted, so you can’t end up in a retry loop.
• It’s capped at 500 contacts. If more than 500 recipients were quality-blocked, the follow-up covers the first 500.
• Paige waits before it looks. WhatsApp reports quality failures asynchronously — a message that looked sent can flip to failed minutes later — so Paige leaves about 2 hours after the original completes before working out who to retarget. The follow-up still sends around 24 hours after the original finished.

The switch is hidden on self-hosted projects, which manage their own recipient list and so can’t be retargeted from Paige.

​

Your contacts and their tags

The Contacts sub-tab is your address book: everyone this project can broadcast to. Each person shows their name, their number, the tags they carry, whether they’re opted in, and whether WhatsApp has confirmed the number is reachable. Search by name or number at the top, and page through 50 at a time. Tags are the point of this screen. A tag is just a label — vip, interested, trial — and tags are what segments use to find people later.

​

Filtering by tag

Click any tag on a contact to see only the people who carry it. The line above the list tells you how many that is, and that number always matches what the same tag reports in the Tags view — it’s the same question asked twice. Clear the filter with the × on the tag chip.

​

Tagging several people at once

Tick the contacts you want, or use Select all on this page, then choose Add tag. Type a name — or pick one you already use from the list underneath — and it’s added to everyone you selected. Their existing tags are kept. You can tag up to 500 people in one go. For more than that, tag them in batches, or ask the Broadcasting agent to tag a whole audience for you.

​

Managing your tags

Switch to the Tags view to see every tag on this project with how many contacts carry it. Click one to jump straight to those people. The pencil renames a tag everywhere it’s used; the trash removes it from every contact that carries it. Deleting a tag never deletes contacts — but any segment filtering on that tag stops matching them, so check your segments afterwards.

​

Two things about tag names

Tags are matched exactly, but capitals in what you type don’t matter. Paige saves every tag in lower case, so Newsletter and newsletter land on the same one. A segment looking for newsletter matches contacts tagged newsletter. Older tags with capitals are listed separately. If contacts were tagged before Paige started saving in lower case, you may see VIP and vip as two rows with two counts — because that is genuinely how they’re stored, and how your segments match them. Paige speaks up when this affects you, showing the count, because tagging the lower-case spelling will not join those people. You can fix it: rename VIP to vip, and Paige tells you it takes two steps — the contacts move to a temporary tag first, then onto vip — and does both for you. Nobody loses the tag along the way. If the second step doesn’t go through, your contacts are safely on the temporary tag and Paige tells you its name so you can finish from the same place.

​

If tagging says the project needs an update

If a message tells you that tagging many contacts at once needs an update for this project, open Settings → Project and click Repair broadcasting, then try again. Sending broadcasts is unaffected in the meantime. If tagging stops part-way through, Paige tells you how many contacts were already updated, so you know exactly what changed.

​

Opting out (STOP) and back in (START)

Your bot handles opt-outs on its own. If someone replies STOP, UNSUBSCRIBE, OPT OUT, or OPTOUT, Paige marks them opted out immediately, before the message reaches your bot. They’re excluded from every broadcast from then on — including ones you’ve already scheduled, because the check happens at send time. The match is case-insensitive and forgiving of extra spaces and punctuation, so “stop”, ” STOP. ” and “Unsubscribe” all work. But the word has to start the message and stand on its own, so “please don’t stop messaging me” won’t trigger it. Your customer gets no confirmation reply when they opt out — they send STOP and hear nothing back. That’s worth knowing if someone asks whether it worked.

​

Coming back

Someone who opted out by mistake, or who later changes their mind, can re-subscribe themselves by replying START — or UNSTOP, SUBSCRIBE, or OPT IN. They’re marked opted in again, they get a short confirmation reply this time, and they’re back in future audiences as normal. The same rules apply: the word has to begin the message and stand alone, so “can we start tomorrow?” and “startup” go to your bot as ordinary messages.

​

Next