A host with fourteen flats in Porto hires someone part-time to cover guest messages in the evenings. The plan lasts four days. The WhatsApp Business app is installed on the host's phone, tied to the host's number, and there is no second seat to give the new person — only the option of logging her in as the host. Two people are now typing into the same thread as the same identity, neither can see which messages the other has already answered, and by Friday a guest has been told two different check-in times by the same business.
Nothing broke. The app did exactly what it is built to do. It was simply never built for two people, and the host had assumed that a business app implied a business team.
The confusion starts with the naming. Meta publishes two different products whose names differ by one word, and they are not tiers of the same thing. Buying the wrong one costs money in one direction and costs evenings in the other.
Two Products That Share a Name
The WhatsApp Business app is an app. You download it, you register a phone number, and it behaves like WhatsApp with a business profile bolted on. It is free. It runs on a phone, with companion devices logged into the same account. One person, one number, one inbox.
The WhatsApp Business Platform — the thing everyone still calls the WhatsApp API — is not an app and has no inbox. It is an interface that software talks to. On its own it displays nothing and does nothing; it is the pipe through which some other system sends and receives messages on a number you control. You never "open" it. You either buy software that has already built on it, or you build that software yourself.
That is the whole distinction, and almost every wrong decision follows from missing it. Asking whether you should "upgrade to the API" is like asking whether you should upgrade from a car to a road. The Platform is the surface other tools drive on, and the question is really whether you need one of those tools.
What the Free App Genuinely Does
It is worth being fair to the app, because a lot of content about this topic exists to make hosts feel inadequate for using it.
The WhatsApp Business app gives you a business profile with an address, hours and a website link. It gives you labels, so you can mark threads as arriving today, awaiting payment, or past guest. It gives you quick replies, which are saved blocks of text you pull up with a keyboard shortcut instead of retyping the WiFi password for the four-hundredth time. It gives you a greeting message for first-time contacts and an away message on a schedule. Details shift between versions and regions, so check what your current build offers rather than trusting any list, including this one.
For a host with one or two properties, that is a real product and it is free. Setting it up properly — a profile that looks like a business, a quick-reply library that covers your actual repeat questions — takes an afternoon and is covered step by step in the guide to setting up WhatsApp Business for a vacation rental. Do that before you consider spending anything.
The Ceilings of the Free App
The app's limits are not defects. They are the shape of a product designed for a shop with one owner. They matter to hosts because hosting breaks all of its assumptions at once.
One number, one business. Everything lives under a single number. A second brand, a second city, or a separate line for cleaners means a separate phone or a separate account, and nothing is shared between them.
One identity, not multiple seats. Companion devices let several people be logged in at once, but they are all logged in as you. There is no assignment, no ownership of a conversation, no record of who replied, no way to stop two people answering the same guest. This is the ceiling most teams hit first, and it arrives the day you hire your first assistant.
Quick replies are typing aids, not automation. A quick reply still requires a human, awake, holding a phone, choosing it. Nothing fires because a guest asked a question. The greeting and away messages do fire on their own, but they are one fixed block of text with no idea who they are talking to — which is why an auto-reply reads as an auto-reply. The difference between that and something that actually answers is the subject of AI replies versus basic auto-replies, and it is a difference guests notice immediately.
No reservation context. The app does not know that +44 7700 900123 is the guest arriving on Thursday at flat 3B. It cannot key anything off a booking, because it has never seen your booking. Nothing can be timed to an arrival, a checkout or a stay length.
Nothing initiates. You cannot arrange for the door code to be sent at 15:00 on arrival day. You can set a reminder for yourself to send it, which is a different thing and fails on the day you are driving — and a door code that arrives late is one of the commonest ways self check-in breaks.
Portability is awkward. Chat history lives in device backups tied to the phone's operating system, and moving numbers or phones is a manual operation with real failure modes. Back up before you touch anything, and verify the restore rather than assuming it.
What Crossing to the Platform Buys
Everything the list above is missing, and one thing it does not mention.
Software on the Platform can hold your reservations, so a message can be timed to a booking rather than to your memory. It can send to many numbers without anyone typing. It supports multiple people with their own logins, assignment and a visible history of who said what. It can run several numbers under one business. And it can answer at 02:00 in a language you do not speak, because a machine is doing the answering rather than a person with a phrasebook.
The thing not on the list is legitimacy. A Platform number carries a verified display name — your business name shows instead of a raw phone number — subject to Meta's business verification process, which generally means a registered entity and documents to match. Meta's requirements and the badge policy have changed more than once; confirm the current rules before you build a plan around them.
What It Costs, Including the Three Costs Nobody Quotes
The sticker cost is the software. The other three are the ones that surprise people.
A provider in the middle. Access to the Platform generally runs through a Business Solution Provider — a partner that holds the connection and the compliance relationship. Some tools resell that transparently, some make you bring your own. Either way you now have two vendors rather than one, and when something breaks it is worth knowing in advance which of them you call.
Template approval. Outside an open conversation window, you cannot send whatever you like. Outbound messages must follow a template submitted to Meta in advance and approved, with variables in defined slots. Approval is usually fast but it is not instant, rejections happen, and you cannot improvise a new outbound message at 22:00 because a boiler failed. The window mechanic that makes templates necessary — a period after the guest's last message during which you can write freely, and outside which you cannot — is worth understanding properly before you commit, and it is covered in the piece on the 24-hour messaging window. Treat the exact duration and the category rules as things to verify against Meta's current documentation, not as settled facts.
Per-conversation pricing. Meta has generally charged for messaging on the Platform by conversation window rather than per message, with different rates by country and by message category, and it has revised that model more than once — including moves toward per-message charging for template sends. Any specific number printed anywhere, this article included, will be stale. Pull the current rate card for the countries your guests actually message from, multiply by your real conversation volume, and only then decide whether the arithmetic works.
There is also the migration itself. Moving an existing number onto the Platform generally means it stops working in the Business app, and coming back is not a trivial reversal. If you can start with a fresh number and keep the old one alive during the changeover, do.
The Comparison, Side by Side
| Criterion | WhatsApp Business app | WhatsApp Business Platform |
|---|---|---|
| What it is | An app you install | An interface software connects to |
| Cost | Free | Software fee plus Meta's per-conversation charges |
| Who replies | A person, on a phone | Software, with people taking over |
| Automation | Greeting and away messages only | Timed, data-driven, conditional |
| Multiple people | One identity on several devices | Real seats, assignment, history |
| Outbound outside the window | Type it yourself | Pre-approved templates only |
| Knows your bookings | No | Only if the software syncs them |
| Setup effort | An afternoon | Days, plus verification |
| Best fit | One to three properties | A portfolio with a team, or night coverage |
Read the table as a fit exercise rather than a scorecard. The right-hand column is not a better version of the left-hand one; it is a different purchase with running costs attached.
Which One You Actually Need
One to three properties: the app. Message volume at that size is a handful of conversations a week, and a person can carry it. Anyone telling a two-property host that they need the API is selling something, and the tell is that the pitch is about capability rather than about how many hours a week the host currently loses. Set the app up properly, build a quick-reply library, and revisit in a year.
Four to ten: the zone where it depends. Property count is the wrong variable here. The useful signals are: you have hired or are about to hire someone to answer messages; you have missed an arrival message and found out from the review; you send the same paragraph so often that you can type it from memory; or your guests arrive from timezones in which you are asleep. Two or more of those and the crossover is probably worth costing out. None of them and you are buying capability you will not use, which is the same mistake in the other direction.
Ten and up, or any size with a team: the Platform. Not because the volume is unmanageable, but because the multi-seat ceiling is hard, and coordination failures at that size are expensive in reviews. If you are at this point and shopping, the buyer's guide to WhatsApp automation tools covers what to compare across vendors, and the pillar on WhatsApp guest communication sets out what the channel is being asked to do in the first place.
One more signal worth taking seriously: if the failures you are having are content failures rather than volume failures — guests asking things your instructions never covered — neither product fixes that. Better instructions do, and they are free.
Honest Limits
The Platform does not make your messaging good. It makes it programmable. A house manual with the wrong door code in it becomes a wrong door code delivered automatically, at scale, in thirty languages. Most hosts who feel their guest communication is failing have a content problem, and the tooling decision is a distraction from it.
Template approval is a genuine constraint on tone, not a formality. You will write a message you like, it will be rejected for a reason you have to guess at, and you will rewrite it flatter. Plan your outbound library early rather than the week you go live.
Pricing on this channel is not stable, and neither are the policy thresholds. Everything in this article about windows, categories and verification is a description of a mechanism, not a quotation of current terms — verify against Meta's documentation before you rely on any of it.
Provider lock-in is real and is rarely discussed at signup. Your number, your approved templates and your conversation history sit with whichever provider holds the connection. Ask what leaving looks like — what exports, how the number is released, how long it takes — before you sign, not after.
And about Welco: it is a WhatsApp AI assistant for hosts, so it lives on the Platform side of this line. If you run two properties, the free app is working, and your evenings are your own, Welco is not solving a problem you currently have, and saying so plainly is more useful than selling you a crossing you do not need. It is also not a property management system — no channel management, no pricing, no payments, no locks. It stops at the conversation.
The Operational Picture
The decision is not app versus API. It is whether repetition and timing have started costing you more than software would, and that is a question about your week rather than about Meta's product line. Most hosts cross over later than the marketing suggests and earlier than their own reluctance suggests.
When the crossing is worth it, Welco answers verified guests on WhatsApp in their own language, syncs reservations by iCal from Airbnb, Booking.com, VRBO or Hostaway without a PMS, and flags the requests that need you — stay extensions, early check-ins, emergencies — with an urgency level. If that sounds like the problem you actually have, request a demo and test it in the simulator before a guest ever sees it.