AI contact center guide
AI Receptionist Pricing Explained: Per-Call, Per-Minute, and What You Actually Pay
How AI receptionist services and software are priced, why the cheapest advertised tier is an anchor rather than a plan, and how to model the bill at your real call volume.
AI receptionist pricing looks simple until you model it against your own volume, at which point the advertised number and the invoice stop resembling each other. The gap is not dishonesty. It is that most of this market inherited its pricing model from human answering services, where every conversation has a real labour cost.
There are only three shapes in the market: metered per call, metered per minute, and a flat platform fee. Once you can tell which one you are looking at, the arithmetic takes about five minutes and it will change which product you buy.
This post walks the three models, does the arithmetic honestly, explains where the underlying costs actually sit, and lists the questions worth asking before you commit a phone number to anything.
The three pricing shapes
Per-call metering charges for each conversation, usually with a small monthly allowance included. It is the dominant model in the answering-service world because it mirrors the provider's own cost structure: a human answered, so a human is billed.
Per-minute metering charges for talk time. It is more common in telephony-native products and it is the more honest unit for voice, since a ninety-second confirmation and a twelve-minute complaint are not the same event. It also introduces an incentive nobody enjoys, which is a system that benefits from longer calls.
A flat platform fee charges for the seat, the brand, or the workspace, and treats conversations as included capacity. It is only viable when the marginal cost of the two-hundredth conversation is close to the cost of the first, which is true for automation and structurally untrue for staffed services.
Why the cheapest advertised tier is an anchor, not a plan
The entry tiers in this category are small on purpose. Published AI receptionist plans commonly start around ten dollars a month for roughly ten calls, then twenty-five dollars for twenty-five calls, forty-five for fifty, and ninety for a hundred. That works out to about a dollar per call at the bottom and roughly ninety cents per call as volume rises.
Ten calls a month is not a business with a phone problem. It is a sole proprietor whose phone barely rings, and that plan exists to occupy the cheapest slot in every comparison article. Nobody who needs an AI receptionist is served by ten calls, and the provider knows it.
Look instead at the tier that matches your actual month. If you take a hundred calls, you are not shopping the ten-dollar plan, you are shopping the ninety-dollar plan, and you should compare that number against what else ninety dollars buys.
Do the arithmetic at your real volume
Take a business handling two hundred conversations in a busy month, which is an ordinary number for a clinic, a trades company, or a small agency in season.
On a metered plan with a fifty-call allowance at forty-five dollars, the next hundred and fifty calls at roughly ninety cents each add about a hundred and thirty-five dollars, landing the month near a hundred and eighty dollars. That covers phone calls only. Texts, web chats, and emails are not included, because in a phone-only product they are not channels at all.
Now add the tools that do cover those channels. Shared-inbox platforms for small business start in the hundreds of dollars per month on annual contracts, and AI reply features are frequently a paid add-on on top. Stack a phone answering product and a messaging platform and the real monthly cost of covering four channels climbs quickly, before anyone has booked an appointment.
- Count conversations in your busiest month, not your average month
- Count them per channel, including the texts and chats you currently ignore
- Price the overage, not the allowance
- Add every module that is not included in the base tier
- Then compare a single flat number against that total
The costs that sit underneath every quote
No matter who you buy from, some costs are physics and cannot be priced away. Telephony is metered by carriers, so inbound minutes and phone numbers have a real per-unit cost that someone is paying. Voice AI consumes speech-to-text, model inference, and text-to-speech per second of conversation. Email and SMS have their own per-message economics.
Vendors handle this in one of two ways. They bundle the usage into the plan and absorb the variance, which is simpler for you and means the plan price carries a margin buffer. Or they pass usage through at cost and charge a platform fee on top, which is more transparent and means your bill moves with your volume.
Neither is wrong, but you should know which one you signed. The failure mode of bundling is a plan that seems expensive in a quiet month. The failure mode of pass-through is a surprise in a busy one. Ask the vendor to show you a real invoice at your volume rather than a pricing page.
What actually moves your bill
Three variables dominate, and only one of them is the one people worry about.
Call length matters more than call count on per-minute plans, and an AI that cannot end a conversation cleanly is a billing problem as much as a customer-experience one. After-hours coverage matters because the calls you were missing were disproportionately after hours, so enabling AI reliably increases the number of conversations you pay for. And spam matters more than anyone expects: unfiltered robocalls and junk web-chat sessions are conversations, and on a metered plan you pay for them.
Ask specifically what happens to a robocall. If the answer is that it is answered, transcribed, and billed like any other call, model that into your busiest month, because it will be there.
Questions to ask before you commit a number
Pricing pages are marketing. These questions get you to the invoice.
The last one is the one people skip, and it is the one that reveals the product's real boundary. If booking an appointment costs extra, then the automation is doing intake and the calendar is still your problem.
- What is my bill at two hundred conversations, itemised?
- Are SMS, web chat, and email included, or is this phone-only?
- Is usage bundled or passed through at cost?
- Do robocalls and spam chats count against my allowance?
- Is there an annual contract, and what happens if I leave?
- Can I sign up today, or do I have to speak to sales first?
- Does the AI book the appointment, or does booking cost extra?
The line items that are easy to miss
Beyond the headline plan, three costs routinely appear on the invoice and almost never on the pricing page.
Contracts are the first. Several platforms in the adjacent messaging category sell on annual terms, which means the monthly figure you compared is a commitment rather than a subscription. Ask what happens if you leave in month three, because that answer is the real price of being wrong.
Modules are the second. Features that sound core are frequently sold separately: AI reply capability as a paid add-on, appointment booking as its own staffed service, additional locations or brands at a per-unit rate. Price the configuration you will actually run, not the tier you will actually buy. The third is seats, where a product priced per user quietly punishes you for letting the whole team see the inbox, which is the entire point of a shared inbox.
- Annual contract, or month to month?
- Which advertised features are paid add-ons?
- What does a second brand, location, or number cost?
- Is it priced per seat, and does that discourage team access?
- What is the cost of leaving, and can I export my history?
How we price, and why
Contact Center HQ charges a flat platform fee: Starter at ninety-nine dollars for a single brand proving AI intake on live channels, Growth at two hundred and ninety-nine for teams running voice, SMS, email, and web chat together, and Operator at seven hundred and ninety-nine for agencies and operators running multiple brands with separate routing.
The reason it is flat rather than per-call is the incentive. We do not want a product that makes you nervous when the phone rings, and we do not want to sell you an appointment-booking upgrade for the thing an AI receptionist should obviously already do. Carrier, AI, and email usage is reported transparently so you can see what the conversations actually cost.
Whether that is the right price for you is a question the arithmetic above answers better than we can. Run it at your busiest month, across every channel, and compare the totals rather than the headlines.