Smith.ai alternatives: the one that stops the meter.

If you are looking for a Smith.ai alternative, the question worth asking is not which competitor is cheaper per call — it is whether you should be buying calls by the unit at all. Every productised answering service works the same way: the meter runs on the phone ringing, and it runs again on the things that make the call worth taking.

We build AI receptionists, so we have an interest in your answer. Everything below about Smith.ai and Ruby comes from their own published pricing, linked at the bottom, so you can check it rather than take our word for it.

Ruby vs Smith.ai: different meters, same idea

One clarification first, because it causes real confusion when people go looking. Smith.ai — frequently written Smith AI, without the full stop — is one specific company, not a category. Searching for a Smith AI alternative turns up a mix of that company’s actual competitors and unrelated AI answering products, and they are not the same thing.

Ruby bills per minute. Smith.ai bills per call. That is the whole difference between them, and it decides only which kind of call hurts. Long conversations cost you at Ruby; high volume costs you at Smith.ai. Neither model changes as your business grows, except upward.

The add-ons are where a Smith.ai bill actually lands

The per-call price buys the call being answered. It does not buy the call being useful. Smith.ai’s own published rate card prices those separately, per call, every time:

What you probably want the call to do	Smith.ai, per call	Built agent
Book the appointment in your calendar	$1.50	Included
Write to an additional CRM	$0.50	Included
Notify you by SMS, Slack or Teams	$0.50	Included
Send a follow-up message afterwards	$0.50	Included
Record and transcribe the call	$0.25	Included

So a call that gets booked, written into your CRM, notified and recorded carries several dollars of add-ons on top of the per-call rate — and carries them again on the next call, and the one after that. The pricing is transparent and it is on their site. It is simply a model that charges you more the better your business does.

Where both of them stop

They answer your phone. They do not become part of how your business runs. The qualification is configured within what the product supports rather than around how you actually decide what is worth your time, and the integrations reach the systems the product has decided to reach.

What a built agent does instead

It is trained on your business and connected directly to your systems, and then it stops costing more every time it works. Your website content, PDFs and FAQs go in, so it knows what you actually do. You agree how it speaks and what it must never say. You define when it escalates to a person and what the edge cases are in your trade. It ships in 25 languages, and it always tells callers it is AI.

Then it books into your live calendar, writes the record into your CRM, sends the notification and stores the transcript — because those are what the agent is for, not extras on an invoice.

Why the setup fee is the point, not the obstacle

A productised service starts cheaply because there is nothing to build. You are added to a platform, given a script, and metered from the first call. That is genuinely faster, and it is why the per-call rate never comes down: the whole cost of serving you sits in the calls themselves.

A build inverts that. Setup starts at £495 and covers the work of making the agent yours — the training, the rules, the integrations, the testing. What that buys is a marginal cost that collapses afterwards: £195 a month including 200 connected minutes, then £0.80 a minute, falling to £0.40 on the largest plan. Nothing extra for booking the appointment, writing to the CRM, notifying you or recording the call.

The arithmetic follows from that. A metered service is cheapest in month one and gets more expensive every month you succeed. A build costs more in month one and then stops moving. The full rate card is published, so you can run your own numbers rather than take ours.

If you are buying from the UK

Both Smith.ai and Ruby are North American businesses that publish in US dollars. That is not a criticism, but it does add two things a sterling buyer carries and rarely prices.

• Currency. A dollar price is an exposure that runs for the life of the contract, not a one-off conversion.
• Hours. Confirm that cover genuinely spans your working day and your out-of-hours rather than a US business day — out-of-hours is usually the reason for buying in the first place.
• Data. Ask where call recordings and transcripts are stored and processed, in writing. If you handle health, financial or children’s data it is a UK GDPR question you will have to answer for yourself eventually.

Questions that expose the difference

Ask these of anyone you are considering, including us. They are the ones where a metered product and a built agent give visibly different answers.

Common questions

What is the best alternative to Smith.ai?

For a business whose calls need to end in a booking or a CRM record, a built AI receptionist is the strongest alternative, because the actions that Smith.ai prices per call are included rather than metered. The other Smith.ai competitors — Ruby, PATLive and similar — solve the same problem the same way Smith.ai does, so switching between them changes the invoice rather than the model.

Is Smith.ai expensive?

Its headline is per call, but the published rate card also charges for the things that make a call useful: $1.50 to book an appointment, $0.50 for a notification, $0.50 per additional CRM and $0.25 to record and transcribe. A call that books an appointment, updates your CRM and is recorded therefore costs meaningfully more than the per-call price suggests, and it costs that again on the next call.

What is the difference between Smith.ai and a built AI receptionist?

Smith.ai is a product you rent per call, configured within what the product supports. A built agent is trained on your own content, style, escalation rules and edge cases, connected directly to your calendar and CRM, and priced as a one-off build plus a monthly fee. The practical difference is that its cost stops climbing with your call volume.

Is Smith.ai or Ruby cheaper?

It depends entirely on your call pattern, because they meter different things — Ruby per minute, so long calls cost more; Smith.ai per call, so volume does. That is the more useful point: whichever you choose, the bill rises every time the phone rings, and the only way to change that is to stop buying calls by the unit.

Do Smith.ai and Ruby work well for UK businesses?

Both are North American and publish in US dollars, so a UK buyer takes on exchange-rate exposure for the life of the contract and needs to confirm that hours of cover span a UK working day rather than a US one. It is also worth asking in writing where call recordings and transcripts are stored and processed, which matters under UK GDPR.

Why does a built agent have a setup fee when Smith.ai does not?

Because something is being built. The setup covers training the agent on your website content, PDFs and FAQs, agreeing how it speaks and what it must never say, defining when it escalates to a person, and handling the edge cases specific to your trade — in 25 languages. A productised service has no agent to train, which is why it starts faster and why it then charges you for every call forever.