What is the cheapest way to try an AI receptionist?

Start with a vertical product in your industry on its lowest tier, or a DIY builder if you enjoy the tinkering. Run it alongside your current process for a month rather than replacing anything. A serious pilot tells you more than any pricing page — and costs less than one week of a custom project.

Are there per-message or per-minute charges to watch for?

Some platforms meter per conversation or per minute of call, which makes the bill grow with your success. Flat-fee products remove that risk but may cap volume. Whatever the model, estimate your busiest month rather than your average one, and check overage pricing before signing — that is where the surprises live.

When is a custom AI receptionist actually justified?

When your call or message flow contains proprietary logic no product implements: unusual booking constraints, in-house systems that must be written to, escalation rules specific to your operation. If a vendor nods along to your description of the flow, buy. If they go quiet, that silence is the business case for building.

// ai

How much does an AI receptionist cost? Three honest answers

4 min read · all guides

Ask a dozen vendors what an AI receptionist costs and you will get a range spanning four orders of magnitude, which is why most pricing articles are useless. The real question is which of the three families of solution fits your situation, because they do not compete on price — they compete on who owns the ongoing work.

This guide breaks down the three tiers with real numbers, shows how to compare total cost of ownership rather than the sticker price, and is direct about when the expensive options are not worth it.

DIY chatbot builders: cheap, if your time is free

Platforms for building your own assistant — visual flow builders, or a generic bot framework wired to a language model — start at tens of dollars per month in software fees. The software is not the cost. Somebody has to write the conversation flows, connect the calendar or booking system, handle the edge cases, and watch the logs when a customer says something unexpected. For a technically confident owner with spare hours, that trade is fair. For everyone else, the "cheap" option quietly becomes the expensive one: we have seen owners spend whole weekends maintaining bots that a flat-fee product would have run without them. Count your hours at your real rate before comparing prices.

Vertical SaaS: the flat-fee middle tier

The second family is purpose-built products for one industry, sold as a flat monthly fee. In restaurants, for example, HousemanAI — housemanai.eu, from the same studio that publishes this guide — handles reservation requests and customer messages on WhatsApp for around €49 a month, with no per-message metering and no flow editor to maintain. The economics work because the vendor amortises one product across many businesses: you get the polished version of the bot you would have half-built yourself. The limits matter too. You adopt the workflows the product supports, and anything beyond its feature set means asking the vendor, not configuring it yourself. For standard operations in a defined industry, this is where the value concentrates.

Custom builds: five figures and up, and worth it only sometimes

A bespoke receptionist — built by a studio, wired into your proprietary systems, with your exact escalation rules — starts at five figures and behaves like any software project: scoped, staged, maintainable. It is the right choice when the phone flow itself is a competitive asset: proprietary booking logic, multi-site scheduling with unusual constraints, integrations with systems no vendor supports. It is the wrong choice for a standard booking flow, where you would pay custom money for a wheel that already exists. Ask one question before deciding: if I describe my reception flow to a product vendor, do they nod or go quiet? A nod means buy, not build. Reserve custom work for the flows where the nod never comes.

Compare TCO, not sticker price

Across all three tiers, the reliable comparison is total cost of ownership over twelve months: subscription fees, your hours, setup costs, and the cost of being wrong — missed bookings, mishandled complaints. A €49 flat fee with zero hours attached often beats a €30 builder subscription plus eight hours a month of your own fiddling; the spreadsheet makes that obvious once time is priced in. Ask vendors three questions before signing: what happens when the conversation goes off-script, who maintains the integrations when the underlying APIs change, and how do I leave if it does not work? The quality of the answers predicts the real bill far better than the price list does.

Frequently asked questions

What is the cheapest way to try an AI receptionist?: Start with a vertical product in your industry on its lowest tier, or a DIY builder if you enjoy the tinkering. Run it alongside your current process for a month rather than replacing anything. A serious pilot tells you more than any pricing page — and costs less than one week of a custom project.
Are there per-message or per-minute charges to watch for?: Some platforms meter per conversation or per minute of call, which makes the bill grow with your success. Flat-fee products remove that risk but may cap volume. Whatever the model, estimate your busiest month rather than your average one, and check overage pricing before signing — that is where the surprises live.
When is a custom AI receptionist actually justified?: When your call or message flow contains proprietary logic no product implements: unusual booking constraints, in-house systems that must be written to, escalation rules specific to your operation. If a vendor nods along to your description of the flow, buy. If they go quiet, that silence is the business case for building.

// next steps

Working on something adjacent? Tell us about it — or read the services we deliver today.