Comparisons
How much does a WhatsApp chatbot cost for a student residence? The cost items, one by one (2026)
Anyone searching for what a WhatsApp chatbot costs tends to find a monthly figure and little else. In a residence, the cost is the sum of different items: what Meta charges for each message, the AI model, onboarding, integrations, keeping the house manual up to date and staff time, which is sometimes saved and sometimes not. Here they are one by one, with who charges for each and what it depends on, so that quotes that look nothing alike can be compared.
The short answer
There is no single price because it is not a single thing. A WhatsApp chatbot for a residence is the sum of six items, and they do not all appear on the same invoice: one is charged by Meta, others by the software provider, and another is put in by the residence itself in staff hours. Comparing two quotes without separating the items means comparing different things.
| Item | Who charges for it or puts it in | What it depends on |
|---|---|---|
| WhatsApp messages | Meta, to the account that sends them | The category of each message, the country of the number that receives it and the volume |
| AI model | The software provider, within its service or separately | How many messages are read, how much text has to be taken into account to reply and which model is used |
| Onboarding | The provider and the residence's staff | The house manual, the org chart, the shifts and the templates that need approval |
| Integrations | The provider | Whether a room rack in Excel, a dashboard or data from another system has to be read or written back |
| Keeping the manual up to date | The residence, the provider or both | How often timetables, menus and rules change over the year |
| Staff time | The residence | How much of what comes in is repetitive and how much needs a person |
What Meta charges: per message, not per conversation
This is the item that causes most confusion, because it has changed twice in a short time. Since 1 July 2025, Meta has charged per delivered message, not per conversation. Many comparisons still talk about 24-hour conversations, and that model no longer applies. Each message carries one category and is charged at most once, according to Meta's page on charges for the WhatsApp Business Platform.
The categories, with examples from a residence
- Utility: messages tied to something the person did or asked for, such as the status of a fault or the notice that a parcel is waiting at the porters' lodge. They cannot be promotional.
- Marketing: anything promotional, such as an open day. Meta also treats as marketing a template that mixes a service notice with a promotion.
- Authentication: one-time codes to verify someone's identity.
- Service: ordinary replies, without a template, inside the 24-hour window that each of the person's messages opens.
The first three are message templates: messages registered and approved in advance, which are the only thing that can be sent to someone who has not written in the last 24 hours. What each message costs varies with the country calling code of the number that receives it, and utility and authentication messages can come down through volume tiers within the month.
Worth knowing
What changed on 1 October 2026. Meta started charging again for two kinds of message it had stopped charging for: service messages, beyond a monthly allowance for each phone number, and utility templates sent inside an open window.
A quote prepared before that date may be based on rules that no longer apply. It is worth asking which of Meta's rules the calculation used.
Who pays Meta depends on how the provider works. If it is a Tech Provider, the residence adds its own payment method, Meta charges the residence for API usage and the provider invoices its own service separately. If it is a Solution Partner, it can extend a credit line to the residence and invoice the platform usage itself. Either way, the residence decides the volume.
In a residence, what moves this item most is the messages that go out without the person having written first: the parcel at the porters' lodge, the fault that has been fixed, the announcement to the whole house. Replies to what a resident asks go inside their window.
The AI model
The assistant needs a language model to understand what it is told, and that model is paid for by usage to whoever runs it, according to the text it processes. What pushes it up: lots of messages, replies that have to take a long manual into account, photos that have to be interpreted and voice notes that have to be transcribed before they can be understood.
The temptation here is to choose the cheapest model, and this is where saving goes worst. A model that understands less misclassifies a fault, gives the wrong timetable or misses that a message is urgent, and every mistake ends up being corrected by a member of staff. The useful thing is to ask which model the provider uses, whether it is included in its service or passed on separately, and whether it can change it without notice.
Onboarding and integrations
This is the item that shows least in a quote and the one that most decides whether the assistant is any use. A generic chatbot gives generic answers; one that works for a residence has to know the residence, and that is prepared once, at the start.
What has to be prepared
- The house manual: house rules, meal and door times, visiting rules, what to do with a parcel and how to cancel a meal. It is the source of everything the assistant answers.
- The org chart: which roles exist and which area each one covers, so that a plumbing fault does not end up in the kitchen. That is the org chart and the shift.
- The shifts: who is on duty at night, at weekends and on bank holidays, and who covers when someone is off.
- The templates: every notice that has to go out while the window is closed is drafted, sent to Meta and waits for approval; according to Meta, the review can take up to 24 hours.
- Registering the number on the WhatsApp Business Platform and, if many people have to be messaged while the window is closed, business verification with Meta.
Integrations
They cost little if the assistant only handles WhatsApp, and more if it has to talk to what the residence already uses. The usual one in a residence is the room rack in Excel: knowing who is sleeping in tonight so their dinner is not counted, and handing the sheet back up to date. Then come the dashboard where staff see the cases and any residence management system that has to be read. Each connection is work, and the quote should say which ones it includes and in what format the data can be taken out at the end.
Worth knowing
The residence's own hours during onboarding do not appear on any invoice, but they are real. Someone has to gather the house rules, explain who covers what and check the first answers. A provider that asks for none of this at the start is offering an assistant that does not know the house.
Keeping the manual up to date
A manual that is not kept up to date ages quickly in a residence. Meal times change during exams, bank holidays come round, the menu changes, the visiting rule changes in summer and the maintenance phone number changes when the contractor does. If the assistant answers with last year's timetable, the cost is not measured in money: residents stop trusting it and go back to phoning reception.
There are three useful questions. Who updates the manual: the residence from a dashboard, or the provider when asked. How long a change takes to reach the answers. And whether that work is part of the service or charged per change.
Staff time: what is saved and what is not
It is the reason a residence considers an assistant in the first place, and it is the item where most unfounded figures circulate. The honest approach is to separate what can be saved from what cannot.
- What can be saved: answering what repeats (timetables, rules, where something is), copying meal cancellations or parcels by hand, and passing each fault from a chat to the person who fixes it.
- What is not saved: the decision on an exception, the conversation with a worried family, an emergency or the repair itself: someone still has to replace the cistern.
- What can backfire: an assistant that passes almost everything it receives to a person, or staff who have to check every answer because they do not trust them.
The serious way to estimate it is to count a real week of messages by type before deciding and, afterwards, to count the same things again in the same way. A savings figure that does not say how many residences it comes from or over how many days is a brochure, not a measurement.
Where Bedel fits
Bedel handles the residence's WhatsApp on the WhatsApp Business Platform and uses templates for what has to arrive while the window is closed: the parcel at the porters' lodge, the status of a fault and announcements. For announcements it respects opt-outs, removes duplicates and does not repeat the same text to the same audience within 10 minutes. It imports the room rack in Excel and hands the same sheet back up to date. In the dashboard, hours saved are counted one by one against a scale of minutes that is on show.
What it does not do: it does not count as saved what staff did by hand, it does not translate broadcasts, and each residence is one account, with no combined view of several. Everything else, feature by feature, is under features.
What to send when asking for a quote
With this list, any provider can give a price that looks like what will actually be paid:
- The type of residence and the number of places, and whether there are international residents and which languages they write in.
- A real week of messages counted by type (questions, faults, meals, parcels, families) and the times they arrive.
- Which features are needed from the start: faults, meals, parcels, announcements, booking of shared spaces.
- The org chart, the night and weekend shifts, and who covers each area.
- How many announcements go out each month and to whom: the whole house, one floor, the families.
- Whether the room rack is in Excel or another system, and which data the assistant would have to read.
- Whether there is already a WhatsApp Business number and whether it should be kept.
- Who will keep the house manual up to date during the year.
Bedel does not publish its prices: the price is discussed on a call, with this list in front of you.
Frequently asked questions
How does Meta charge for WhatsApp messages?
Per delivered message since 1 July 2025, not per conversation. What each message costs depends on its category (marketing, utility, authentication or service) and on the country calling code of the number that receives it, and utility and authentication messages come down through volume tiers. Since 1 October 2026, Meta has also charged for service messages beyond a monthly allowance per number. The amounts are on its page for developers.
Why are quotes for a WhatsApp chatbot so different from one another?
Because each one includes different items. One counts only the software and leaves Meta's messages out; another includes the AI model while a third passes it on separately; one comes with onboarding and the house manual, another hands over an empty tool that has to be filled in. To compare them, put the six items in columns and see what each quote covers and what it leaves out.
Who pays Meta, the residence or the provider?
It depends on the provider's role. With a Tech Provider, the residence adds its own payment method and Meta charges the residence for API usage, while the provider invoices its own service separately. A Solution Partner can extend a credit line to the residence and invoice the platform usage itself. The quote should state in writing which of the two models applies and whose name the account is in.
Is it cheaper to answer with the WhatsApp Business app?
On messages, yes: Meta does not charge per message for what is sent from the app. In exchange, whoever has the phone in front of them answers, there is no routing by role and no record beyond the chats, and every reply is written by a person. For a small residence with low volume it can pay off. When there are shifts, several areas and night-time messages, the cost moves to staff hours.
What pushes up the cost of the AI model?
The volume of messages, the length of the manual the model has to take into account to reply, the photos it interprets and the voice notes it transcribes. Choosing a weaker model to save money usually turns out expensive: it misclassifies, gives wrong answers and means a person has to correct it. It is worth asking which model is used, whether it is included in the service and whether it can be changed without notice.
How much does Bedel cost?
Bedel does not publish its prices. The price depends on the size of the residence, the features switched on and the volume of messages, and it is discussed on a call. For it to be a useful price from day one, bring the list on this page: places, a week of messages counted by type, features, org chart, shifts, announcements and whether the room rack is in Excel.
Comparisons
- Student accommodation software: alternatives (2026)The five types of tool a student residence combines, seven alternatives according to their own websites, and Bedel with what it does and what it does not.
- WhatsApp group or assistant for a student residenceSix ways to run a residence's communication channel, compared by what they solve, privacy, night-time, who answers and the record they keep.
- WhatsApp Business or API: which suits a student residenceThe WhatsApp Business app and the WhatsApp Business Platform compared point by point, with what changes for a student residence.
For each type of centre
- Colegios mayoresFor colegios mayores: the dining hall with its cancellations, packed lunches and allergens; activities with their call and their survey; families and applicants; and what belongs to management, to management.
- University residencesFor residences owned by or affiliated with a university: support in each resident's language, repairs and cleaning with outside contractors, the arrivals of each semester and enquiries about places.
- Student residencesFor student residences and PBSA operators: WhatsApp support in the resident's language, repairs with an owner and a deadline, arrivals, presence and parcels.
- ColivingFor coliving and flex living: support in each resident's language, repairs with an owner and a deadline, common-room bookings, parcels and announcements without groups.
- Researcher residencesFor researcher, graduate and visiting-faculty residences: arrivals any week of the year, replies in each person's language, contracted maintenance and cleaning, and the end of every stay.
- Hostels and alberguesFor hostels, albergues and pilgrim hostels: house questions answered in each guest's language, repairs with an owner and a deadline, and the night covered with nobody at the desk.
- Holiday apartmentsFor holiday apartments, aparthotels and mid-stay lets: replies in the guest's language, repairs with a technician and a deadline, housekeeping requests between stays and house rules.
- Build to rentFor build-to-rent and flex living buildings: repairs with an owner and a deadline even when maintenance is outsourced, parcels with a notice, shared-space bookings, building notices and house rules answered.