Three billing models dominate AI answering service pricing, and they answer three different questions. Per-minute plans price the time the system is on the call. Per-call plans price the call itself, regardless of length. Flat plans price the month, regardless of either. The right model for your shop depends on how your call volume behaves across the year — and how much variance in that volume you're willing to absorb.
This page walks through each pricing model honestly, shows how the math works at realistic garage door call volumes, and lays out the fair-use and "unlimited" clauses that change what the headline number actually means.
| Model | What the meter measures | What makes the bill rise | Best fit for |
|---|---|---|---|
| Per-minute | Time on the call | Long calls, holds, transfers, after-hours multipliers | Low call volume, short calls |
| Per-call | Each answered call | Call count | Predictable volume, variable call length |
| Flat / unlimited | The month | Nothing | Spiky volume, after-hours demand, growth |
Each model has a place. The mistake is treating them as interchangeable on a rate card. They aren't.
Per-minute plans charge for the time the AI spends on the call — including triage, hold time, transfers, and any silence on the line. A commonly seen planning figure in this category is around $1 to $2 per minute, often with a small base fee and an after-hours multiplier on top.
How the math plays out for a garage door shop:
Where per-minute pricing falls apart for this trade is exactly where the calls get valuable. A trapped-car call at 11 PM that needs a careful, two-minute triage, a service window offered, a confirmation, and a clean handoff is a longer call than a daytime opener quote. The longer, more thoughtful call — the call that books a high-revenue job — costs more on a per-minute plan than the cheap inquiry.
For the failure modes that show up in invoices, see per-minute pricing pitfalls.
Per-call plans charge a fixed amount per answered call, usually in the range of about $1 to $2.50 a call. Call length stops mattering. Call count does.
How the math plays out for a garage door shop:
Per-call is more predictable than per-minute for shops with variable call length, but the bill still tracks volume. Storm weeks and spring season still cost more. The meter has just changed what it counts.
The trade-off is that providers using per-call pricing sometimes shape the call experience to fit the meter — shorter calls, less thorough triage, faster handoffs. A garage door shop that wants careful booking and clear triage on every call should ask how the billing model affects call handling. For the mechanics and trade-offs, see per-call pricing explained.
A flat rate AI answering plan prices the month, not the call or the minute. One number, every month, regardless of how many calls came in or how long they ran. Ava is in this bucket: $97 first month, then $297/month flat, unlimited calls, no contract, cancel anytime.
How the math plays out for a garage door shop:
Flat rate is the only model where the worst month and the best month cost the same. For a trade with spiky volume — storm weeks, cold snaps, and the spring rush — that's the math you want. The surge is where you make money; the surge shouldn't also be where your answering bill punishes you for growing.
A flat rate AI answering plan is also the cleanest line item on the budget. There's no surprise invoice at the end of the month. The number is the number. For what "unlimited calls" really means in the fine print, see "unlimited calls": what it really means.
Flat rate AI answering pricing relies on a word that vendors love: unlimited. Read what it means before you sign.
Honest flat-rate plans have a fair-use clause that protects against abuse — a competitor calling 200 times a day to harvest your triage script, for example. The clause usually says calls must be genuine customer calls in normal volume, and gives the vendor a way to talk to you if usage looks abnormal. That's reasonable.
Less honest flat-rate plans use "unlimited" loosely:
The simplest test: ask the vendor, in writing, what happens to your bill in a month with 3× your normal call volume. If the answer is a clear "the bill doesn't change," the flat plan is honest. If the answer involves a calculator, it isn't.
For the contract language to watch for, see unlimited calls with no contract: what that means for you.
These are planning figures, not quotes. Pull your own call counts and rerun the math:
| Monthly call volume | Per-minute (~$1.50/min, 4-min avg) | Per-call (~$1.75/call) | Flat (Ava, $297/mo) |
|---|---|---|---|
| 80 calls | ~$520 | ~$140 | $297 |
| 150 calls | ~$950 | ~$263 | $297 |
| 250 calls | ~$1,550 | ~$438 | $297 |
| 400 calls (storm month) | ~$2,450 | ~$700 | $297 |
The pattern is the same in every direction. Per-minute is the most expensive option at any meaningful volume. Per-call is the cheapest at low volume and climbs with you. Flat is the most expensive at very low volume and the cheapest at anything approaching real-world call counts.
The crossover point depends on the per-call rate and the average call length, but a reasonable planning figure: if your shop is averaging more than about 100 calls a month, flat rate AI answering will usually beat per-minute, and often beats per-call too. The moment call volume gets spiky — which it does, every spring — flat pulls further ahead.
Billing model doesn't just change the invoice. It changes the experience:
The honest answer is that billing model and call quality are connected, and the cheapest rate card isn't always the cheapest answer.
The trade has a few features that push the answer in a particular direction:
Those four features together push toward flat-rate AI answering pricing for almost every shop in this trade. The flat plan absorbs the spiky volume, removes the after-hours meter, lets the system triage carefully without watching a clock, and aligns the vendor's incentive with the shop's.
The exception: a very low-volume shop — under about 50 calls a month with mostly short calls — can sometimes come out ahead on a per-call plan. Run the math either way before you sign.
When one provider quotes a flat $297 a month and another quotes a per-minute plan at $1.50 with a small base, the comparison isn't the headline numbers. It's the projected total at your volume.
Three steps:
Then compare against the revenue from one or two captured calls at your average ticket. The plan that costs more than the revenue it captures isn't a deal, no matter what the rate card says.
AI answering service pricing is a question of structure, not sticker price. A per-minute plan punishes long calls. A per-call plan punishes high volume. A flat-rate plan punishes no one — the bill is the bill. For a garage door shop with spiky volume, after-hours demand, and a calendar that needs filling, the flat-rate model is the cleanest fit: predictable cost, full coverage, and an incentive structure that lines up with the shop's own.
Ava's offer — $97 first month, then $297/month flat, unlimited calls, no contract, cancel anytime, with a 30-day "First 10 Leads On Us" guarantee — sits in the flat-rate bucket and lets the shop judge the service by the leads it captures, not by the math on the page.
Pull your call logs. Pick your model. Then have Ava call you and hear the difference.
Call the live demo and have Ava call you now — hear exactly what your customers will hear when they call your shop.