Hotel Chatbot Pricing: How Much Does a Hotel Chatbot Cost in 2026?

A hotel chatbot costs from $349 a month on our own price list, and from 1 October 2026 Meta charges for WhatsApp replies on top.

Those are the two lines on the bill, and only one of them appears on most pricing pages. The subscription is the visible number: ours is published in full - $349 a month for a single boutique property, $699 for the full platform on every channel, from $1,490 a month per property for groups (read 29 September 2026). Meta's message fees are the other, they vary by the guest's country, and from 1 October 2026 they cover something they did not cover before. You can work out your own share of them in our WhatsApp pricing calculator for hotels, which carries Meta's list rates for 17 markets.

First, the payment-method deadline: 30 September 2026

If your WhatsApp Business account has no payment method on file by 30 September 2026, Meta will deliver your service messages inside the free monthly tier and stop delivering them once that tier is used (Meta's WhatsApp pricing documentation, read 28 September 2026).

Read that as an operational failure rather than a billing one. Replies simply stop, part-way through a month, with nothing the guest can see. No error, no bounce, no warning on their side - a question sent to the hotel that never gets answered, and a guest who concludes nobody is there. If you run WhatsApp for a property, check that a payment method is on the account before 30 September. It takes minutes, and it is the cheapest item on this page.

What else changes on 1 October 2026

Two charging changes take effect, both from Meta's non-template message pricing documentation (read 28 September 2026):

• Service messages become chargeable per message. These are the free-form replies you send inside the 24-hour customer service window - the answer to "what time is check-in?", the confirmation that the late arrival is noted. They had not been charged since November 2024.
• Utility messages sent in response to a guest inside an open window become chargeable too. They had been free since July 2025.

Four things shape what it actually costs:

• A free tier of 1,000 delivered service messages per month, per business phone number. Meta's wording on its WhatsApp pricing page (read 28 September 2026) is: "Each business phone number has one shared free tier of 1,000 delivered service messages per month." It resets monthly with no roll-over, and one-to-one and group deliveries draw on the same pool.
• Service rates match the utility and authentication rate for that market. No premium for replying.
• Rates are per delivered message, set by the guest's country, not the hotel's. On Meta's rate card effective 1 October 2026 (read 18 September 2026) that is $0.0053 for an Israeli number, $0.022 for a UK one and $0.055 for a German one - more than a tenfold spread across three markets a single property might serve in one morning.
• There are no volume tiers for service messages. Meta continues to offer volume tiers for utility and authentication messages, so a property sending a lot of templates still earns a lower unit rate on those. Replies are simply priced flat.

The practical consequence is that guest messaging now has a marginal cost per reply, and the cost depends on who your guests are. Two properties of the same size, one hosting mostly Israeli guests and one mostly German, have very different WhatsApp bills for identical behaviour. Our calculator takes your own guest mix and estimates Meta's share; it covers 17 markets and is built from Meta's published card.

The three things that set the platform price

Conversation volume. This is the big one, and it is structural. Meta charges per delivered message and language-model inference is charged per token, so both scale directly with how many guests message you and how much they say. A vendor's costs go up with your volume, so their price does too - either transparently, through tiers, or invisibly, through caps and overage in the small print.

Properties and channels. Per-property pricing is standard for groups. Each additional channel - WhatsApp, Instagram, Messenger, web chat, Telegram - usually adds cost, because each is a separate integration to maintain and, for WhatsApp, a separate metered service. If you run several properties, how many WhatsApp numbers you run is a pricing decision as much as a routing one, which one assistant or many across venues works through alongside the free service-message tier that comes with each number.

Integration depth. A read-only PMS connection that surfaces reservation details is cheap. Two-way write-back that records requests, attaches documents and updates the guest profile is not, because it is more engineering and more per-PMS work. Some PMS vendors charge their own fee for the API connection, separately from anything the chatbot vendor charges.

Features - the things listed on the pricing page - move the price less than any of these three.

What a real quote looks like

We publish our own numbers, so use them as the shape of a quote rather than as the market. A single boutique property with modest volume sits on an entry tier; a property wanting the full guest journey across every channel with two-way PMS sync sits a tier up; groups are quoted per property with a floor. Add-ons such as an extra channel or an additional PMS connection are priced per property, per month. That is all on the pricing page, which is the test worth applying to anyone else: if the numbers are not published, ask why, and ask what they are at your volume.

Elsewhere in the category, some vendors publish and many do not, and the figures that circulate in roundups are frequently out of date. We have not verified anyone else's current prices first-hand, so this page does not quote them. What we can tell you is what to ask.

The question that reveals the real price

What does this cost at my actual monthly conversation count?

Count your inbound guest messages for a week, across every channel. Multiply by four. Ask for the price at that number. Then ask what happens at double, because a working deployment grows volume - guests message more once they learn they get an instant reply, and a channel that used to be quiet becomes the main one.

From October, ask a second question alongside it: at that volume, how much of it is service messages, and who pays for the ones beyond the first thousand? A vendor who answers both in one sentence has thought about it. One who needs to get back to you has not been asked before, which is its own signal.

The costs that are not on the invoice

Four regularly surprise properties in the second month.

Meta's message fees. Charged per delivered message at a rate set by the guest's country; some vendors bundle them, some pass them through. Ask which, and ask what the mark-up is if they bundle. Our WhatsApp pricing calculator estimates Meta's own share for your guest mix, which is the number a bundled quote should be compared against.

PMS integration setup. Sometimes a one-off charge from the chatbot vendor, sometimes from the PMS vendor, sometimes both. Ask who invoices what.

Template approval time. Every proactive message on WhatsApp - pre-arrival, check-in link, departure reminder - is a template that needs approval. This costs days rather than money, but it lands during onboarding, and a vendor with a library of approved patterns saves you most of it.

Someone to own it. No vendor invoices this and every deployment needs it: a person at the property who keeps the knowledge base current and watches the escalations. An hour a week at a small property. Budget it, because an assistant with no owner drifts out of date within a quarter.

What is worth paying for

Three things justify a higher tier, and they are the same three that separate products in a demo.

Two-way PMS write-back, because it is the difference between answering questions and doing work.

Escalation to the channel staff actually watch, with a fallback if nobody responds, because the unanswered request at 2am is the one guests remember.

Multilingual including voice, because the guest writing in a language nobody on the team speaks is the one currently getting the worst service.

One more has been added by the October change: resolving a guest's question in fewer messages now has a price attached. An assistant that answers in one clear reply costs less to run than one that needs four, and the gap widens with every guest. That is a cost argument, not only a service one.

Insist on a pilot

A short pilot at real volume before a contract is the only way to see any of this with your own guests. Not a demo - a demo shows the happy path. A pilot shows the escalation, the write-back, and what a Polish voice note actually gets back.

A vendor confident in the product will agree. One who will not has told you what you needed to know.

Further reading

• Hotel chatbot or AI guest assistant - what a hotel chatbot is, and what separates the scripted bots from an assistant worth paying for.
• WhatsApp pricing calculator for hotels - Meta's charges for your own guest mix, 17 markets, from 1 October 2026.
• How to simplify guest communication across booking channels - the two kinds of platform, and which problem each one solves.
• WhatsApp replies to hotel guests stop being free in October - the change in full, with what stays free.

A note on the free allowance, added 5 October 2026. The monthly allowance of 1,000 delivered service messages per business phone number is quoted above as it was read on Meta's pricing page on 18 September 2026. Re-reading both of Meta's pricing pages on 5 October 2026, that allowance is no longer stated on either of them. What those pages do still say, and what this piece rests on, is that from 1 October 2026 service messages are charged per message. Meta may have removed the allowance, changed it, or moved it somewhere we have not found. Until it can be seen again, treat the figure as unconfirmed and check your own billing in WhatsApp Manager rather than this number.