Migrating from AiSensy without losing your WhatsApp number
The number moves through the BSP process, not a CSV import. Templates, approvals, opt-outs and the 48-hour cutover checklist.
The number is the hard part. Contacts and templates can be exported; a WhatsApp Business number has to be migrated through the BSP process without losing the display name or the template approvals.
Start with an inventory: contacts, suppressed numbers, active templates, automations, API keys and webhook consumers. Freeze new campaign changes 24 hours before cutover.
Move in this order: verify the new BSP, submit templates, migrate contacts with opt-out state, rebuild the two highest-volume automations, test with a seed list, then switch the number. Keep the old BSP read-only for 72 hours.
The most common failure is not technical. It is sending a campaign before the opt-out list is imported. Do not do that.
What this means for your team
Every action has a trace with tool, latency and cost.
Every consent has a notice version and a timestamp.
Every escalation has a named human and an audit entry.
Every number has a denominator and a window.
If a number in this article is not defensible, it does not ship. That is the editorial standard.
Contacts and template text are the easy part. The WhatsApp number, the opt-out list and the flow logic are the migration.
Do not cancel the old BSP before the new number is verified. The number is the dependency everything else waits on.
Import suppression state before the first campaign, not after. A send to unsubscribed contacts is the most common self-inflicted damage.
Templates have to be resubmitted with matching categories and variables, so keep the old template IDs for the transition.
LoopNow Starter is ₹2,999 a month plus 18% GST. AiSensy records a lower WhatsApp-only entry price, and for a single-channel broadcast business that matters.
A 48-hour migration is realistic for the data. Budget longer for rewriting the rules your old flows encode.
What moves in two days and what does not
Migration guides tend to describe the export. The export is a CSV. The migration is the set of decisions that were never written down: which contact counts as opted in, which template belongs to which journey, which branch handles the exception that happens twice a week. Those decisions live inside the old tool, and the new tool cannot import them, because they were never data.
The mechanical work is genuinely fast. Contacts, opt-out state, template text, automations, API keys and the WhatsApp Business number can all move, and LoopNow quotes about 48 hours for a typical migration done for you. That estimate covers the data movement and the cutover, not the redesign of a flow that grew over two years. Read the two numbers separately or you will plan a two-day project and run a two-week one.
Figure 1. The sequence that works. Number verification sits early because every later step depends on the new business account being live.
The useful way to plan is to split the work into three lists: data that moves, configuration that is rebuilt, and knowledge that has to be re-established. The first list is short. The second is a day of work. The third is the reason migrations feel slow, and it is the only part where a template cannot help you.
The number is the migration
A WhatsApp Business number is tied to a business account and a solution provider. Moving between providers is a supported process, but it is not a database update. The display name and the template approvals travel with the account, and the migration completes through Meta’s process. If you cancel the old provider first, you can end up with a number that is attached to nothing while the new request is pending, which is the one outcome nobody wants.
Start the number work on day one and treat everything else as parallel. Confirm with both providers what happens to the display name, how long the number is unavailable, and what evidence Meta needs. Put the cutover in a low-traffic window, and tell the team who owns customer communication if the number is briefly offline. A WhatsApp channel that goes quiet without warning generates more support load than a planned outage.
Figure 2. The number path, which gates the rest of the plan. The old provider stays active until the new number is verified.
The number also carries your quality rating and your messaging limits. A brand-new business account starts without history, so the first campaigns after migration should be small and to engaged contacts. Do not migrate on a Friday and send a fifty-thousand-message broadcast on Saturday. Give the account a week of normal, good-quality traffic before you test its ceiling.
Contacts, opt-outs and the list you must not mail
Export contacts with their opt-out state, not just their phone numbers. A phone number without consent is a liability, and a suppression list that stays behind is worse than no migration at all. Import the suppression list before the first campaign, then verify it by counting: the number of suppressed contacts in the new tool should match the number in the export, allowing for duplicates.
Duplicates are the second problem. Indian contact lists accumulate the same person under a country code, without one, with a leading zero and with a name spelled three ways. Deduplicate on the phone number in E.164 format before import, decide which record wins, and merge consent state conservatively. If one record says opted out and another says opted in, the safe merge is opted out until the person says otherwise.
Figure 3. The import pipeline. Normalise the number and carry the suppression state before anything is segmented or mailed.
Tags and attributes need a mapping table, not a best guess. Write the old field, the new field, the transformation and the owner. Anything that does not map cleanly should go into a notes field rather than be dropped, because the sales team will ask where the lead source went three weeks later. A mapping table is also the document you hand to the person who runs the next migration.
Templates: categories, variables and re-approval
Templates do not transfer as approved objects between business accounts in every case, and the safest assumption is that they need to be resubmitted. The category has to match the intent. A payment reminder submitted as marketing will be rejected or approved at the wrong rate; a promotional offer submitted as utility will be rejected when a reviewer reads it. Match the old category exactly unless the content genuinely changed.
Variables are the other rejection source. A template with a variable in the header, a variable in the body and a button parameter needs a sample value for each, and the sample has to be plausible. Reviewers reject templates with blank samples, placeholder text or variables that could carry promotional content. Keep the old template IDs in a spreadsheet next to the new ones, because the transition period will have both in flight.
Figure 4. The four fields that decide whether a resubmitted template is approved. Each one needs a real sample, not a placeholder.
Approval takes time and is not entirely in your control. Submit early, submit in batches, and keep the rejection reasons. A rejected template is information: it tells you how the reviewer read the category. If you are migrating a large library, expect to retire some templates rather than port all of them. The ones that have not sent in six months are usually not worth the review cycle.
Rebuilding chatbot flows as workflows
A chatbot flow is a tree. The customer walks a branch, the branch ends, and the conversation either resolves or falls to a human. A workflow is closer to a process: it can wait for an event, call a tool, check a policy and pause for approval. Moving from one to the other is not a copy operation, because the tree encodes the old model’s limits as much as the business rules.
The productive approach is to write the rules down before opening the new builder. For each flow, list the trigger, the questions, the decisions, the tools, the exceptions and the handoff. Then rebuild only the two or three flows that carry the most volume or the most revenue. Everything else goes on a backlog with a date. Teams that rebuild all twenty flows before cutover usually delay the cutover and ship the same bugs in a new interface.
Figure 5. Two automation models. The workflow is more powerful and more work; only rebuild the flows where the difference is worth it.
The rebuild is also a chance to fix what the tree forced you to accept. A flow that asks for an order number before it checks whether the caller is logged in can be replaced by a lookup. A flow that escalates every refund can be replaced by a policy check with a human gate above a threshold. Measure the old flow’s completion rate first, so the new one has a baseline rather than an opinion.
Webhooks, API keys and the messages in flight
The cutover has a state problem: messages sent by the old provider may still be delivered after the new one takes over, and their status webhooks will arrive at the old endpoint. If the new system never sees those events, your reports will show messages as sent but not delivered, and the support team will not know which conversations are open. Plan for a reconciliation window rather than a clean switch.
Keep the old webhook consumer alive for at least 72 hours, write the events into a table with the source marked, and reconcile the two streams after the cutover. On the new side, verify the signature on every webhook, make the handler idempotent on the message ID, and return a success quickly rather than doing slow work inline. A webhook endpoint that times out causes retries, and retries cause duplicate replies.
Figure 6. The cutover reconciliation path. Both streams write to one log so the team can see what the old provider delivered after the switch.
API keys deserve their own line in the runbook. Rotate them at cutover, store them in a secret manager rather than a spreadsheet, and check which internal services still hold the old key. A key that keeps working after the migration is a quiet dependency, and it will fail at the worst possible moment, usually during a campaign.
What LoopNow adds, and what it costs
LoopNow runs WhatsApp on the official Business API, with email, voice, workflow and GEO on the same runtime. In the comparison data we publish, WhatsApp is included rather than being the whole product, email is included, voice covers fourteen Indian languages, and the workflow builder supports human gates. Those are the differences that justify a broader platform, and they are also the differences you pay for.
The entry price is ₹2,999 a month plus 18% GST. For that you get 5,000 marketing contacts, 25,000 emails, 1,000 WhatsApp messages, 500 voice minutes and 1M AI tokens, with three agents and five seats. The WhatsApp allowance grows to 5,000 messages on Growth at ₹7,999 and 20,000 on Scale at ₹24,999. Overage is published: ₹1.15 per marketing message and ₹0.22 per utility or authentication message.
Figure 7. Included marketing contacts per plan. Suppressed and unsubscribed contacts are not counted, so the import does not inflate the bill.
If you only send WhatsApp broadcasts, none of that is a reason to move. The comparison data records a lower WhatsApp-only entry price for AiSensy, and a single-channel business should take that seriously. The case for LoopNow starts when the second channel already exists in a separate tool, or when the support queue ends in a phone call that nobody can afford to keep routing by hand.
The price comparison, with GST on both sides
Price comparisons fail when one side is quoted with tax and the other without. LoopNow publishes prices exclusive of GST, so a ₹2,999 plan is ₹3,538.82 on the invoice at 18%. AiSensy bills in INR and issues GST invoices, so its headline price also needs the tax added before you compare. Whatever the two numbers are, compare them at the same volume, with the same tax treatment and the same twelve-month term.
Then add the costs that are not on either price list. A second vendor for email, a calling vendor for the follow-up, the engineering hours to keep consent state in sync, and the finance time to reconcile three invoices. Those costs exist whether or not they appear on a comparison page, and they are the reason a broader platform sometimes wins despite a higher headline price. They are also the reason it sometimes loses.
Figure 8. The question that decides most migrations. If the second channel is not already running, consolidation is a future cost, not a present saving.
The honest test is a spreadsheet with four columns: your current stack, the current stack plus the channels you are about to buy, LoopNow at your volume, and the migration cost amortised over twelve months. If the difference is small, choose the simpler operating model. If it is large, choose the cheaper one and revisit in a year. Price is a fact; the rest is a preference.
When you should stay on AiSensy
Stay if WhatsApp is the whole plan. A coaching institute sending admission reminders, a clinic confirming appointments, a retailer running a weekly offer to regulars: these are broadcast-shaped businesses, and a focused WhatsApp platform serves them at a price a broader runtime does not match. Stay if the team is small and marketing-owned, because fewer concepts means faster training and fewer ways to misconfigure a send.
Stay, also, if your only complaint is the bill. Saving a few hundred rupees a month is not worth a cutover, a retraining week and the risk of a quiet channel during the transition. And stay if nobody has asked for a second channel, a voice call or a consent export in the past year. Those requests are the real signal that the current tool has become the constraint; without them, a migration is a change of scenery.
Figure 9. The end of the evaluation, in two columns. Both outcomes are legitimate; the mistake is migrating without a reason.
The comparison data lists three situations where AiSensy wins: a single channel with a simple bot, the lowest entry price for that channel, and no need for voice, email or a consent ledger. We wrote those rows because they are true. A migration page that cannot tell you when to stay is a sales page with a timeline graphic.
Rollback, and the 72-hour read-only window
A rollback plan is not pessimism. It is what lets you cut over on a Tuesday instead of waiting for a quiet month that never arrives. The plan has three parts: the old provider stays reachable and read-only for 72 hours, the new provider is tested on a seed list before any real traffic, and one person has the authority to stop the migration without a meeting.
Read-only means you can still see the conversations and export the history, not that you can send. Sending from both providers on the same number is not possible, and attempting it creates duplicate replies and confused customers. Keep the old account for evidence and for the small chance that the number migration has to be reversed, and accept that a reversal is a project rather than a button.
Figure 10. The cutover dashboard. Every row is a thing that can fail silently, which is why it is checked rather than assumed.
After 72 hours, do a post-cutover audit. Compare message counts, delivery rates and template performance against the old baseline. Check that opt-outs from the new tool reach the CRM and that the support team can see the full history. Then archive the old exports, revoke the old keys and write down what you would do differently. The next migration in the company will use that document.
A 48-hour cutover checklist
The checklist assumes one owner and a decision-maker who is available on the day. It also assumes the number migration has already been requested, because that step has an external dependency and cannot be compressed into the final afternoon. If the number is not verified, stop and reschedule rather than migrating the data into a workspace that cannot send.
Two checks are worth repeating because they are the ones teams skip under time pressure. Count the suppressed contacts in the new tool and compare with the export. Then send the seed campaign to a group that includes an opted-out number and confirm it is not delivered. A suppression list that exists but is not enforced is the same as no list.
Figure 11. The cutover sequence in order. The negative suppression test near the end is the one that catches a broken import.
Number verified on the new BSP, display name confirmed in writing.
Contacts imported in E.164 with consent state and suppression counts reconciled.
Top templates resubmitted with matching categories and valid samples.
Two highest-volume flows rebuilt and dry-run on real cases.
Seed campaign sent, including an opted-out number as a negative test.
Old BSP read-only for 72 hours; rollback owner named.
Post-cutover audit scheduled with a baseline to compare against.
What we have verified, and what to confirm
Our comparison rows for AiSensy were checked on 4 November 2026 and are published with the date so you can see how current they are. They cover the channel surface, the billing currency, the entry price position and the situations where a single-channel tool wins. They do not cover AiSensy’s export format in enough detail for us to grade it, and we mark that as not verified rather than assume the worst.
Before you sign anything, ask both providers the same five questions. What happens to the display name during the number migration? Which templates need re-approval, and who submits them? How long is the number unavailable? What is the export format for contacts, flows and conversation history? Who is the named support contact on cutover day? The answers are more useful than any feature table, including ours.
Figure 12. Six questions to put in writing before cutover. None of them is answered by a feature comparison.
If the answers are vague, that is the finding. A migration is a coordination problem between two vendors who have no incentive to make it easy for each other. The team that asks for specifics in writing, sets a low-traffic window and keeps a rollback path will land it. The team that trusts a 48-hour promise without a plan will spend the weekend in a group chat.
Frequently asked questions
Can I keep my WhatsApp number when I leave AiSensy?
Yes, in the normal case, because the number belongs to your WhatsApp Business account and moves between solution providers through Meta’s process. The display name and template approvals travel with the account. Do not cancel the old provider before the new number is verified, and confirm the downtime window with both vendors in writing before you schedule the cutover.
How long does the migration actually take?
LoopNow quotes about 48 hours for contacts, templates, automations, suppression lists and the number, done for you. That covers the data and the cutover. Add design time for rewriting the rules inside your chatbot flows, which is not a copy operation. A team with three live flows should budget a week end to end and keep the old provider read-only for 72 hours afterwards.
Do my templates need to be re-approved?
Assume yes. Templates are tied to the business account and the category, and the safe plan is to resubmit them with matching categories, variables and plausible sample values. Keep the old template IDs next to the new ones during the transition. Submit early and in batches, and retire templates that have not sent in six months rather than paying the review cycle for them.
What happens to my chatbot flows?
They do not transfer as executable objects. Contacts, templates and suppression lists move; a chatbot tree has to be rebuilt as a workflow. Treat that as design work: write down the trigger, questions, decisions, tools, exceptions and handoff for each flow, then rebuild the two or three that carry the most volume. Everything else goes on a dated backlog.
Is LoopNow cheaper than AiSensy?
Not for WhatsApp alone. Our comparison data records a lower WhatsApp-only entry price for AiSensy, while LoopNow Starter is ₹2,999 a month plus 18% GST. The gap narrows when the LoopNow plan replaces a second vendor for email or voice, because those channels and their allowances are already inside the plan. Price both options at your twelve-month volume with tax included.
What is included in LoopNow Starter?
5,000 marketing contacts, 25,000 emails, 1,000 WhatsApp messages, 500 voice minutes, 1M AI tokens, three agents, three workflows and five seats, at ₹2,999 a month plus 18% GST. Overage is published per meter, with marketing WhatsApp messages at ₹1.15 and utility or authentication at ₹0.22. The spend cap is on by default, so overage does not arrive as a surprise.
Will I lose my conversation history?
Your ability to read old conversations depends on what the old provider exports and for how long it keeps the data. Export the history before you cut over, store it with the contact identifier, and check the format is readable without the old application. LoopNow can import contacts and templates, but a full transcript archive is usually a separate export rather than part of the standard migration.
How do I avoid messaging people who opted out?
Import the suppression list before the first campaign and enforce it at send time, not just at import. Reconcile the count against the export, then run a negative test: include an opted-out number in a seed campaign and confirm nothing is delivered. Keep one suppression state per contact across channels, so an opt-out from WhatsApp is respected by email as well.
Should I migrate if I only send broadcasts?
Probably not. If WhatsApp is your only channel, your bot is simple and the entry price is the constraint, a focused WhatsApp platform is the better buy, and our own comparison says so. Revisit when a second channel is genuinely running, when support calls need a voice agent, or when someone asks for a consent export you cannot produce.
