If you run a small business with a phone number, you are getting pitched AI phone answering right now. The demos are impressive, and most of them are real. The hard part isn’t deciding whether this works. It’s deciding what you want it to do, because that single choice moves the price more than anything else.
The spread is enormous. At one end, fifty dollars a month gets you something that books appointments and answers the same five questions all day. At the other end, a capable agent connected to your systems can handle the large majority of your inbound calls, including the complicated ones. Both are sold as “AI answering.” They are not remotely the same product.
The Simple End: Answer and Book
The entry-level version picks up, answers common questions from information you give it, and books an appointment into your calendar. Hours, directions, availability, pricing, “are you open on Sunday.”
For a lot of businesses that is genuinely enough. If most of your calls are people asking the same handful of things and trying to get on the schedule, this end of the market solves your actual problem for the price of a phone line. It starts around fifty dollars a month.
What it won’t do is know anything. It can’t look up what a customer ordered last time, can’t see that today’s crew is running two hours behind, and can’t move a booking that lives in a system it isn’t connected to.
The Capable End: Agents That Know Your Business
Further up, an agent is wired into the systems that hold your real information: your booking platform, your point of sale, your customer records. That changes what it can do on a call.
It can tell a caller what they ordered last time and reorder it. It can move an appointment and send the confirmation. It can tell someone their part came in, because it can see that it came in. It can take the entire call from hello to booked without a person touching it, and it can do that on twenty calls at once at 6pm on a Friday.
Deployed well and given time to improve, agents at this end handle something like 80 percent of inbound calls without a human. The remaining fraction gets handed to a person, with everything the agent already collected, so the caller doesn’t start over.
That capability is what you’re paying for, and the price reflects it. The question worth asking yourself before you shop is which end of that range your business actually needs.
Yes, Including the Hard Calls
The old advice was that AI handles the easy calls and humans take the difficult ones. That line is moving fast, and it has moved further than most people realize.
The better agents now handle complicated, multi-step calls, and they handle upset callers well. Sometimes better than a person does. An agent doesn’t get defensive when someone is shouting, doesn’t take it personally, and doesn’t get short at the end of a long shift. It stays level, keeps asking the right questions, and gets to the fix. Anyone who has been the person answering the phone at the end of a bad week knows how hard that is to do consistently.
That doesn’t mean everything should go to it. There are calls you want to take yourself because the relationship matters, or because a customer who has been with you fifteen years deserves your voice. That’s a business decision, not a technical limit. Make it deliberately, rather than assuming the software can’t cope.
The Three Costs Behind Any Quote
Wherever you land in that range, every quote bundles three separate costs into one number, which is why comparing two of them feels impossible. Pull them apart and it gets easy.
The platform. The monthly subscription: the phone system or contact center seats, the place where the agent is configured, the reporting. Priced per user or as a flat fee.
The conversations. What the agent costs while it’s actually talking, usually billed per minute or per conversation. This scales with how busy you are, and it’s the part people forget to model.
The integration. Connecting the agent to your booking system, point of sale and customer records. This is the line that separates the fifty-dollar product from the capable one, and it’s where two similar-looking quotes diverge most. One vendor includes the connection to your booking platform. The other assumes you’ll pay for it separately or do without.
There’s a fourth cost that never appears on a quote and is never zero: somebody has to keep the agent current when your hours, prices or services change. Budget a couple of hours a month of someone’s attention, and more in the first month.
The Two Numbers That Decide It
Forget headcount replacement. At a dozen employees you are not eliminating a position, and any vendor framing it that way is selling to the wrong business.
It comes down to two numbers you can work out this week.
How many calls do you miss? Not calls that ring. Calls nobody answers, or that go to voicemail and never get a callback. Your phone system can tell you, and so can your cell carrier if that’s where calls land. Most owners guess low by half.
What is a call worth? The average value of what a call turns into, a booking, a table, a job, a repeat order, multiplied by the share of calls that convert.
Multiply those together and you’ll know within a few minutes whether you’re shopping at the fifty-dollar end, the capable end, or not shopping at all. Do that arithmetic before you sit through another demo.
Where It Goes Wrong
Buying capacity you don’t use. Enterprise contact center platforms are priced for enterprise volume, with seat minimums and annual terms. A twelve-person business on a plan built for eighty agents is paying for a call center it will never run. Ask what the smallest workable configuration is, not what the recommended one is.
Going live before the integration is real. This is the expensive mistake. An agent that sounds perfect but can’t actually see your calendar will confidently offer a customer a slot you don’t have. That costs you more than the missed call would have. Until it’s been tested against your real data, an agent should answer questions, not make promises.
What to Ask For in the Quote
Ask for these as separate lines, in writing:
- Platform fee and seat count, with the minimum seats spelled out.
- Per-minute or per-conversation rate, and what counts as a conversation. A caller who hangs up after five seconds shouldn’t cost what a five-minute booking costs.
- Integration cost, by system, named specifically: your booking platform, your point of sale, your customer records.
- What share of calls they expect it to handle without a person, and what that was for a business like yours.
- Contract length and what happens to your phone numbers if you leave. Numbers should be portable. Get it in writing.
- Where call recordings and transcripts are stored, for how long, and whether they’re used to train anything. This matters more if you take payment details or health information over the phone.
- A pilot option. After-hours and overflow only, for a defined period, before it answers everything.
Start With After Hours
The lowest-risk way in is the calls you’re already losing: after hours, and overflow when every line is busy. Nothing that works today gets disturbed, the agent handles calls that currently reach nobody, and after a month you have your own numbers instead of a projection.
That’s a much easier decision to reverse than replacing your main line on day one.
Where I Come In
I work with providers that build and run AI voice agents and contact center platforms, at both ends of that range. They do the deployment, the integration and the ongoing support. My part is matching you with the one that fits what you actually want the agent to do, and reading the quote with you before you sign it.
What I look for is whether the three costs are separated, whether the integration to your real systems is included or assumed, and whether what they’re selling matches what you need. A business that just wants appointments booked should not be buying a contact center platform, and a business with real call volume shouldn’t be handed a fifty-dollar answering bot.
If you’re weighing one of these now, send me the quote. Working out which end of the range fits your business takes one conversation, and it’s worth having before a demo talks you into a plan built for somebody else.