Keep your WhatsApp Business number when you switch support tools

The number moves with the account. The checklist beside it covers what the old support tool still holds.

A WhatsApp Business number is portable. Meta moves it between providers at the account level, into a new WhatsApp Business Account (WABA), and the identity customers recognize travels with it: the display name, the number's quality rating, its messaging limits and its Official Business Account status. That part of a tool switch is the safe part.

The exposed part is everything else. A support team works in five assets that sit beside the number and do not share its protection: chats, contacts, templates, automations and analytics. Templates are the only one of the five that the migration documents list as carried, and then only the high-quality approved ones. The other four live mainly inside the support tool being left.

The timing matters as well. Meta bills the WhatsApp Business Platform across four message categories (marketing, utility, authentication and service), and Blueticks reports that the billing logic behind them has been restructured more than once, which gives any support team a reason to re-evaluate its tool. I think a switch made for cost reasons deserves the same discipline as one made for features. Inventory the five assets. Decide what finished means. Keep a way back until you get there.

What moves with a WhatsApp Business number when you change providers?

Meta moves the number together with its display name, quality rating, messaging limits, Official Business Account status and high-quality approved templates. Lower-quality, rejected and pending templates stay behind.

Before trusting any migration promise, sort each asset you depend on with three questions:

1. Does Meta hold this asset, or does the support tool hold it?
2. Does the new provider's migration document list it by name?
3. If it is listed, does it arrive unchanged, reset, or under review again?

The mechanics explain why those questions matter. A WhatsApp Business Account (WABA) is the Meta-side container that holds a number, its templates and its reputation. A Business Solution Provider (BSP) is the Meta-authorized intermediary that connects that account to software. Twilio's migration documentation describes the move as an account-level operation: the customer selects the existing Meta Business Portfolio, creates a new WABA with the new provider, and Meta then migrates the number and its data into it. The support tool never carries the number. Meta does.

The left column follows respond.io's help article on moving a number from another provider, which names each carried item and adds that open conversation windows are retained with the timer not reset. The right column is where the detail sits. Templates migrate by duplication: Meta copies high-quality templates into the new account with the same status, then reviews them against its template categorization guidelines, so some may be set to REJECTED. Template quality ratings do not travel at all. Every duplicated template shows an UNKNOWN rating for the first 24 hours, after which Meta assigns a new one if enough data exists.

The common assumption is that the number is the fragile part of a switch. The documents point the other way: the number, its limits and its Blue Tick are the most protected assets in the move, while the templates behind every outbound message are the least protected. That matters because of how the channel works. A customer's message opens a 24-hour customer service window, free-form replies can be sent only inside it, and outside it a business must use an approved template. A template lost to re-review is therefore a follow-up you cannot send.

Neither migration document describes a way to preview which templates Meta will copy before the number moves. I read that as a reason to audit templates by status and quality before anything else is scheduled. Set side by side, the 4 sources behind this section show one pattern: nothing confirmed as carried lives inside a support tool.

Reviews at Zazachat are written and edited by our editorial team, led by Daniel Calloway, and if the choice of destination is still open, our ranking of the best live chat software is where to compare candidates before a number is released.

Now look at what neither column contains. Conversation history, contact records, bots, routing rules and reports appear in neither list, and that absence is the second migration.

Do you lose chat history, contacts, automations and analytics when you switch WhatsApp support tools?

Plan as if you do. Chat history, contacts, automations and analytics live in the support tool, and neither migration document lists any of them among the assets Meta moves.

The carried list is short and specific. Everything missing from it is most of what a support team touches in a working day. To be clear about the evidence: the two migration documents discussed above name account-level assets only, and nothing in the evidence behind this article shows any provider moving conversation archives, contact records, bots or reports automatically. One of the documents adds a condition that makes the point sharper: message history and uploaded media do not migrate when the source account sits on the On-Premises API.

History is context. In 2023, a customer experience trends report found that 70% of customers expected support teams to have full context, while 60% had dealt with agents who had little or none. A switch that strands the archive in the old tool manufactures that gap on the first morning. Contacts carry a second obligation, because businesses must collect opt-ins before initiating conversations, so an export without the opt-in record is an export you cannot safely message from.

The more durable arrangement ties every channel to a single customer record held outside the inbox, so the history follows the person, not the channel. One sales messaging vendor describes syncing transcripts and summaries to a CRM for that reason, and the logic holds for support. A transcript that already sits in your CRM does not depend on the inbox you are leaving.

Automations are rebuilt, not moved. One enterprise provider's own onboarding description makes the order plain: connect the WhatsApp Business Account, then configure routing and queues, then set up AI agents. Connecting is quick. Configuring is the work. The same goes for reporting, since delivery rate, first response time, resolution rate and escalation rate are all computed inside whichever tool handles the conversation. Your trend lines stop at the cutover unless a baseline was saved.

Does staying on the same provider change any of this? Only the account-level half. Some tools connect to WhatsApp through the customer's own provider account, which means the provider relationship, and the number registered under it, belongs to the business and not to the tool. In that arrangement the first thing to ask is whether the number needs to move at all. The four assets in the table are untouched by that convenience, because all four still sit in the product you are leaving.

How complete an export turns out to be varies by product, and the evidence here does not measure it. I would not accept a vendor's assurance in place of a test file. If the destination is a full help desk, read the Zazachat guide to help desk software with live chat with one question for each product: what does it let you export, and in what format?

That settles what to capture. The order in which you capture it, and what must be true before the number is released, is a separate question.

How do you migrate a WhatsApp Business number without an outage or a surprise bill?

Export from the old tool first, clear Meta's prerequisites second, release the number third, and keep the old subscription live until your checks pass.

With the account-level and tool-level assets separated, the two lists become one ordered plan. It has three phases, and the first is the longest.

Before the cutover

1. Export chats, contacts and reports from the old tool, then open the files to confirm they are complete.
2. Document every automation by trigger, condition and action so it can be rebuilt and retested.
3. Audit templates by status and quality, and resolve pending ones, because pending, rejected and low-quality templates are not carried.
4. Confirm admin access to the verified Meta business account linked to the current provider, with an approved display name and no name change pending.
5. Turn off two-step verification for the number in WhatsApp Manager. If you cannot, the current provider must.
6. Check that the source and destination accounts share a data localization region.
7. Confirm someone can receive the 6-digit code by SMS or phone call on the number.

Steps four to seven are where moves stall. Twilio's migration documentation warns that selecting a different Meta Business Portfolio at registration results in an error, that a region mismatch makes registration fail, and that a change to the destination region can take up to 15 minutes to take effect. None of those checks is difficult. Each one can stop a cutover that was announced to customers.

During the cutover

The transfer itself is brief when the prerequisites hold. One provider's help article says its flow completes within 5 minutes with no downtime, and that messages sent to the number during the move are received once it completes. That is a provider describing its own flow. The failure paths are slower: a code that cannot be sent means retrying after an hour, and retrying a failed migration too quickly means waiting at least a day. A failed attempt can also leave the number showing as already registered, and the fix is to turn off two-step verification and register again.

Pause scheduled broadcasts and automated template sends in the old tool before the release, so no campaign straddles two providers. Schedule the release for your quietest hours, and tell customers beforehand on the channels you already use:

• The number is not changing.
• Replies may be slower during a stated window.
• Which other channel to use if the matter is urgent.

After the cutover

Nothing in the evidence describes a way to send a number back to the previous provider. The rollback window is therefore not a button. It is the period in which the old tool, its exports and its subscription stay intact while three checks run:

1. Send one message from every template you depend on, and note any that Meta has set to REJECTED.
2. Trigger each rebuilt automation with a test conversation.
3. Compare the first reports against the saved baseline.

One provider advises cancelling the old subscription after a successful migration to prevent future charges. I would define successful as all three checks passed, not as the number showing as registered.

What does the switch do to your bill?

Write down the exact time of the migration, so you can check both providers' invoices against it. Messages sent before it are charged by the previous provider, including any delivered afterward, and messages sent after it are charged by the new one. Expect one month with two invoices.

The larger change sits underneath both providers. Blueticks lists three dated changes to Meta's billing: service conversations became free and unlimited on November 1, 2024, charging moved to each delivered template message on July 1, 2025, and regional rates were adjusted on January 1, 2026. A provider pricing guide adds a fourth date, 1 Oct 2026, from which free-form service replies are chargeable beyond 1,000 free service messages per phone number each month.

The provider's own fee sits on top. The same analysis puts the typical markup at $0.003 to $0.010 per message for larger providers, and one platform vendor claims some rivals mark up Meta's charges by 30-60%. The implication is plain. A support team's reply volume is now a cost line, and the tool you move to decides how visible that line is.

Smaller teams still choosing a destination can start with the Zazachat shortlist of live chat for small business, then put the same billing question to every candidate. Ask it in one sentence: are Meta's fees passed through at cost, and where does that show on the invoice?

What should you do before you release a WhatsApp Business number?

Inventory chats, contacts, templates, automations and analytics, export what the old tool holds, and keep that tool running until templates, automations and reports all pass their checks in the new one.

The number cannot be rehearsed on a test line: one provider's 2025 help article says WhatsApp-issued test numbers cannot be migrated. Meta carries it, with its rating and its limits, and a transfer that one provider describes in minutes will only become more routine. I expect the number move to turn into an ordinary onboarding step, which shifts the real work to the two places this article has pointed at: a template audit before the move, and a tool-level inventory that no migration document performs for you.

Cost will keep prompting these decisions. Meta has changed its billing more than once, and each change is a reason to compare tools again. Zazachat is not a WhatsApp provider or a live chat vendor. It is an independent publication that reviews and ranks customer support software, which is why the advice here favors no destination.

Start with one export. Pull a full conversation archive from your current tool, open it, and search it for a customer you remember.