← All guides

Answering Service Pricing Explained (and Where the Traps Are)

Three models, one phone bill

Every answering service price reduces to one of three models. Per-minute plans sell bundles of talk time — Ruby's published tiers, for example, run $250 for 50 minutes up to $1,725 for 500. Per-call plans sell answered calls, like Smith.ai's human plans from $300 for 30 calls. Flat or per-agent plans, common among AI providers, sell a seat or an agent for a fixed fee with a caller allowance, like Goodcall's $79 for up to 100 unique callers. None of these is inherently fairer than the others. Each is a bet about your call pattern — and the vendor, who has seen thousands of businesses like yours, has priced the bet with better information than you have. Your job is to close that information gap before signing.

Why bundles are sized to tempt you

Bundle pricing borrows its psychology from mobile phone plans. The entry bundle looks cheap because it is sized for a quieter business than yours — 50 minutes is about twelve four-minute calls a month. Most growing businesses land one or two tiers up within a quarter, and vendors know the distribution: the advertised 'from' price is real for almost nobody who stays a customer. The honest way to shop is to find your actual minutes first. Your current phone bill, VoIP dashboard or even a week of manual tallies will give you calls per month and average length; multiply them and shop for that number, not for the tier you wish described you. Buying the bundle below your real usage is not thrift — it is choosing the most expensive rate on the page, the overage rate, for a large share of your calls.

Overage: where the money is

Overage rates are the least advertised and most important numbers in this market. On human services they are steep: Ruby's run roughly $3.30 to $5.40 per minute beyond bundle, and Smith.ai's human plans charge up to $11.50 per extra call. Ten unbudgeted minutes a day at five dollars is $1,500 a month — more than many entire AI plans cost in a year. AI services have gentler overages, typically well under a dollar per extra call or a modest per-minute rate past a bundle, but the structure is identical: a cheap base, a metered excess. Whatever you trial, set a calendar reminder to check usage at the end of week one, not month three. Vendors will usually move you up a tier without argument; they are quieter about refunding three months of overage you never noticed.

Per-call plans and the short-call question

Per-call pricing has a hidden definition problem: what counts as a call? A hang-up after ten seconds, a wrong number, a telemarketer, the same customer ringing three times in an hour — on some plans, all four are billable events, and on others the provider filters some of them out. At three dollars a call the difference is trivial; at ten-plus dollars, it is your invoice. Ask for the plan's exact definition, its spam and hang-up policy, and a sample invoice if you can get one. Per-minute plans have the mirror-image problem: rounding. If minutes are billed in whole-minute increments, a 61-second call costs two minutes, and a business taking many short calls can see its effective rate rise by a third.

Setup fees, add-ons and the rest of the invoice

The subscription is rarely the whole bill. Common extras include one-off setup or onboarding fees, per-booking charges (Smith.ai has charged around $1.50 per appointment booked), per-lead or per-message fees, charges for patching calls through to your team, holiday surcharges on human services, and the underlying phone line or number costs an AI agent sits on top of. Individually these are small; together they routinely add 15 to 30 percent to the headline price. None of this makes any particular vendor dishonest — it makes the headline price an unreliable comparison tool. Ask every finalist for a single number: at my call volume, in my busiest recent month, what would the total invoice have been? Then compare those numbers, not the pricing-page banners.

Price your worst month, then decide

The method this page exists to teach is simple: take your busiest month of the last year, not your average one. Run its calls and minutes through each candidate's model, including the overage that month would trigger. The calculator below automates exactly that across the main AI providers and a human benchmark, with every assumption printed on the page so you can argue with it. What you are looking for is not the cheapest average month — it is the plan whose worst month you can shrug at, because that is the month when the phones ring off the hook, every call is revenue, and a pricing surprise is the last thing you need.

Next step

Model your worst month with the AI Receptionist Cost Calculator — free, runs in your browser, no email required.

Advertisement

Sponsored link