How much does an AI receptionist cost? A real breakdown

A chat-based AI receptionist usually costs a flat 50 to 450 EUR a month. A voice-based one usually costs a smaller base fee plus a per-minute rate, which means the headline price tells you almost nothing. The two models fail differently: a flat fee gets expensive when you are quiet, and a per-minute fee gets expensive exactly when you are busy. Most comparison articles list both in the same column and never mention this, which is why the numbers never match the invoice.

The two pricing models, and why they are not comparable

Almost every product in this category prices one of two ways.

Per-minute, for voice. You pay a base subscription, then a rate for every minute the assistant spends on the phone. A voice call costs the vendor real money in speech recognition, model inference and telephony, and those costs scale with time on the line.

Flat monthly, for chat. You pay one subscription for a tier, usually banded by conversation volume or by which channels you connect. Text costs the vendor far less per interaction and does not scale with call length, because there is no call.

The consequence is that they invert under load. A salon taking 40 calls a month is cheap on per-minute pricing and probably overpaying on a flat plan. The same salon in December, taking 300 calls, is the other way round. If you are comparing quotes, work out your busiest month rather than your average one.

What Meta charges on top, which nobody quotes

WhatsApp moved from conversation-based billing to per-message billing on 1 July 2025. Under the current model each delivered template message is billed on its own, priced by category (marketing, utility, authentication), by the recipient country, and by your volume tier. That is Meta's own documentation, not a vendor summary.

Replies inside the window are free. Meta's pricing documentation states that non-template messages are free, and that utility templates delivered inside an open customer service window are also free. When a customer messages you first, a 24-hour window opens, and answering inside it costs nothing at the platform layer.

You pay to start the conversation, not to have it. The cost lands when you message someone outside that window and have to use a template. A booking confirmation sent two minutes after the customer wrote to you is free. A reminder sent 26 hours later is a template message, and it is billed.

A click-to-WhatsApp entry point opens 72 free hours. If the customer arrives from a click-to-WhatsApp ad or a page button, messages including templates are free for the following 72 hours.

The practical version: a chat receptionist that answers people on the day they wrote to you carries almost no platform message cost. The cost appears when your flow depends on chasing people later. That is a design decision in the product you buy, so ask a vendor directly how many template messages their standard flow sends, and how they charge you for them - some bill per message, others sell a package.

Instagram works differently. There is a 24-hour window to respond to a user, confirmed both in Meta's Messenger Platform policy and in the Instagram Messaging API documentation, and inside it you are not paying Meta per message.

A worked example

Take a hair salon. Roughly 250 enquiries a month, most arriving as Instagram messages in the evening, some on WhatsApp. About 60 of them go unanswered until the next day, and the owner reckons a quarter of those are lost. The numbers are illustrative, not a customer case.

On a flat chat plan at the mid tier, the cost is the subscription. Replies go out inside the 24-hour window, so platform message fees stay near zero. The bill is the same in December as in February.

On per-minute voice, that volume never touches the phone at all. The enquiries are text. A voice product answers a channel these customers are not using, and the per-minute rate applies to the calls it does take, which is the smaller half of the problem.

That is the real decision: not which product is cheaper, but which channel your customers actually message you on. Paying anything to automate a channel your customers have abandoned is the most expensive option on the table.

What Avenos costs, plainly

Since this is our own site, here are our numbers rather than a vague contact us.

The WhatsApp module is a message package, not a per-message pass-through: you buy a bundle, and every WhatsApp message we send on your behalf comes out of it, including the ones Meta does not charge us for. We would rather give you a number you can plan around than a bill that moves. Instagram carries no such fee and no module.

Where this is the wrong purchase. If most of your bookings arrive by phone call, buy a voice product instead. If you take fewer than about 30 enquiries a month, answer them yourself, because at that volume no subscription pays for itself. And if your problem is that you have too few enquiries rather than too many, this is not the tool you need.

The costs that do not appear on pricing pages

Setup time. Connecting channels and defining services, durations and rules is usually an hour or two of your time. Ask whether that is self-serve or a paid onboarding.

Calendar integration. If a tool cannot write directly into the calendar you already use, someone is retyping appointments, and that is the job you were trying to remove.

Reminders. This is where template message costs actually appear, because a reminder sent the day before an appointment is by definition outside the service window. Ask how reminders are sent and what they cost.

The exit. If you leave, do your customer conversations and upcoming bookings come with you? Ask before you sign, not after.

The short version

Compare like for like, because a flat monthly fee against a per-minute rate is a comparison of two different products. Work out your busiest month rather than your average. Check where your enquiries actually arrive. And if a vendor cannot tell you plainly what platform fees ride on top of their subscription, treat that as the answer.

For the difference between an assistant, a chatbot and a booking page, there is AI receptionist vs chatbot vs booking software, and the flow itself is on how it works.