Predictable AI voice agent pricing: a flat plan, not a per-minute meter

An AI voice agent's price decides whether your bill is predictable or a lottery, and it all comes down to one detail: how they charge you for it. Per-minute billing works like a taxi meter that starts running the moment someone picks up; a flat monthly plan works like a fee you already know before the first phone rings. A flat plan with an included allowance gives you predictability; the meter gives you a scare at the end of the month. Here's why the per-minute model has a broken incentive, how the pool of minutes and messages actually works, and why the spend cap is what turns a high-volume month into something calm.

The meter bias: why per-minute billing is broken

Most voice services work like a taxi: the meter runs from the moment someone picks up. Every minute on the phone costs you, no matter the reason. It sounds fair—you pay for what you use—until you realize that model spikes precisely when things are going well for you.

The better your business does, the more it costs you. You launch a campaign, your ad goes out, high season arrives… and the meter speeds up. The problem isn't just that it goes up: it's that you don't know how much it'll go up. Budgeting with a meter is guessing, and guessing wrong at the end of the month is exactly what a small business can't afford.

We think price should be the boring part of the equation. You should be able to focus on serving people well, not on watching a counter that starts from zero on the first of every month.

Flat monthly plan: an included allowance and a price you already know

With Totem you pay a flat monthly plan with an included allowance of minutes and messages. Usage draws down that allowance: the minutes the AI spends handling calls and the messages that go out count against what you've already signed up for. You know the number before the month starts, not after.

The difference from the meter is fundamental, not a nuance. With a per-minute counter, the world decides your bill; with a flat plan, you decide it when you pick your plan. A high-volume month is still predictable because you're not watching a counter climb from zero, but drawing down an allowance you already paid for.

How minutes and messages are counted

For the model to be honest, it helps to be clear about what draws down your allowance. It's short and there's no fine print.

Minutes are, simply, the time the AI spends handling calls. That time counts the same whether it takes the call or makes it: what's measured is the conversation, not the direction. And it's counted by the second, not rounded up to the next minute—a 1:57 call bills 1:57, not 2:00—so you never overpay for rounding. WhatsApp templates have their own cost, but that charge is set by Meta for starting the conversation; it's not a Totem add-on. For the fine detail of how minutes, messages, and templates are counted, read the guide on how minute, message, and template usage works, and top-ups.

Spikes and caps: why a high-volume month won't scare you

The reasonable question is: "okay, flat plan, but what if I blow through it one month?" That's what two pieces working together are for.

First, top-ups: if you run short on your allowance in a given month, you top up a one-time balance at a transparent rate, with no drama and no penalties. The extra rates are flat and the same across all plans—for example €0.02 per extra message—except the voice minute, which drops with the plan (€0.22 on Voice, €0.20 on Pro); and if you need one more teammate or an extra booking page, those are €5/mo and €10/mo respectively, with calendars always free and unlimited. Second, the spend cap: you set a limit and the platform won't go past it without warning you. Before your bill gets close to a number you didn't want to see, the system stops and tells you. There's no silent counter running in the background.

The underlying idea: if a spike is a one-off, a top-up covers it; if what really changes is your stage—more channels, one agent becoming several orchestrated agents, more people on the team, or subteams within a family—then you move up a plan for capacity, not just for balance, and you're back to the flat fee that fits you. The system nudges you toward predictable pricing, not toward the meter.

Which businesses gain the most from predictable pricing

Any business with phone support gains from a bill that doesn't depend on the month's volume, but the effect is strongest where campaigns and high seasons create clear spikes.

• Med spas and dental practices. An ad that works brings an avalanche of calls and messages. With a flat plan, that spike is a calendar filling up inside a fee you already knew. According to our clients in the sector, handling it this way translates into up to +50% booked appointments and −35% no-shows when reminders are also turned on.
• Hotels. High season multiplies direct bookings in several languages and at any hour; with a flat plan, your budget doesn't depend on how many come in.
• Real estate. The portal lead wants an answer in minutes and arrives in bursts; budgeting per minute during those bursts is impossible.
• Restaurants and salons. High volume of short conversations to book. The meter would make them hard to predict; with the included allowance, they're a flat fee.

In all of them the pattern repeats: the per-minute model turns volume into uncertainty, and uncertainty is exactly what a small business doesn't want in its budget. If the opportunity-cost math interests you more than the per-minute math, see why your AI voice agent costs less than one missed call a day. And if you want the exact breakdown of each plan, you'll find it in how much an AI voice agent costs per month.

All of this lives inside the same monthly plan, with no lock-in and live in 48–72 hours. And watch for one key nuance: the plan isn't chosen by volume. Because top-ups absorb usage spikes, what really determines your plan is capacity and complexity (chat only, voice, a single agent, several orchestrated agents, or a family with subteams plus a knowledge base and RAG) and your team size (users, calendars, calls in parallel). Consumption is secondary. If you're still not sure which plan fits your stage, read which TotemAI plan to choose: Chat, Voice, Pro, Studio, or Scale.