Zoom AI receptionist pricing: the 100-minute catch

Zoom just did the one thing that unsticks most AI receptionist deals: it stopped asking you to replace your phone system. On July 9 it released a standalone version of its Virtual Agent Receptionist that runs on any existing business phone

Zoom just did the one thing that unsticks most AI receptionist deals: it stopped asking you to replace your phone system. On July 9 it released a standalone version of its Virtual Agent Receptionist that runs on any existing business phone setup, priced at $29.99 a month for 100 minutes. Zoom AI receptionist pricing is now a line item, not a telephony migration.

Then read the second half of that sentence again. One hundred minutes.

That is the whole story, and none of the coverage I read bothered with the arithmetic.

Zoom unbundled the part that actually mattered

Every AI receptionist conversation we have with a business owner hits the same wall around minute six. They like the idea. They do not like being told the first step is migrating their phone system.

That objection has killed more good deployments than price ever has. A dental practice with three lines, a front desk that knows the handset by feel, and a number printed on a decade of paperwork is not going to rip that out to test an AI. Fair enough.

So the interesting sentence in Zoom's announcement is not the price. It is Chris Moss, GM of Zoom Phone, saying businesses "shouldn't have to replace their phone system to benefit from AI." Zoom detached the receptionist from Zoom Phone entirely. It now sits on top of whatever has a dial tone in your back office.

Good product decision. It also puts real pressure on the specialist vendors whose pitch was "we work with your existing number." That is not a differentiator anymore.

The price is $0.30 a minute, and that is the top of the market

Divide $29.99 by 100 and you get just under $0.30 per minute. Annual billing at $24.99 brings it to just under $0.25.

Now compare that to what the infrastructure costs. Vapi charges $0.05 a minute to host the call and passes model costs through at cost, or free if you bring your own API keys. Retell publishes $0.07 a minute for voice infrastructure, and its own cost breakdown puts a realistic all-in setup between $0.13 and $0.31 a minute once you add speech-to-text, an LLM, text-to-speech, and telephony.

So Zoom's $0.30 lands at the ceiling of that range. You are paying a full retail margin for someone else to have already assembled the stack, which is a completely legitimate thing to buy. We break down where each platform's money actually goes in the real cost per minute across Retell, Vapi, and Bland if you want the component-level version.

The rate isn't the problem. The cap is.

What 100 minutes actually buys

Here is arithmetic, not a statistic. I'm not going to hand you an "industry average call length," because yours is the only one that matters and it is sitting on your last phone bill.

• At two minutes a call: 100 minutes = 50 calls a month.
• At three minutes: 33 calls a month.
• At four minutes: 25 calls a month.

Across roughly 22 business days, the middle number is about one and a half calls per day. A one-location plumber does that before lunch. The six physiotherapy practices we rolled AI receptionists out to would have burned through the entry plan in the first week, and we have the no-show and after-hours numbers from that rollout to show where those minutes went.

A receptionist that runs out of minutes on the 12th is not a receptionist. It is a voicemail with better manners.

— Arthur, Bunny Honey Club

There is a structural oddity in flat-rate minute bundles worth naming. The businesses that need an AI receptionist most are the ones drowning in inbound calls. Those are exactly the businesses the entry tier fits worst. Priced this way, the product is cheapest for the customer who needs it least.

The overage rate is the number Zoom did not publish

I went through the newsroom post looking for what happens at minute 101. It isn't there.

No per-minute overage rate, no hard stop, no auto-upgrade tier. It may be published elsewhere in Zoom's pricing pages or handled in-product. But it is not in the announcement, so I won't guess at it in either direction.

That question decides whether this is a $30 product or a $300 product for you. It is the only one that matters at signup.

Where Zoom genuinely wins

Being honest about this: for a real slice of businesses, Zoom is the correct purchase and we would say so.

If you take under 100 inbound minutes a month, want zero engineering, and would rather buy from a vendor your bookkeeper already recognises on the invoice, the case is strong. Ten-plus languages at the entry tier is more than several specialist vendors ship. Live transcription, appointment scheduling, intelligent routing, and human handoff are all in the box. There is a free trial.

That beats a custom build for a low-volume practice, and it flatly beats what most small businesses run now: nobody answering after 5pm.

We are a vendor here. We build and run these for clients, so treat our read accordingly and check the trial yourself.

Where it loses: the moment the call has to do something

Answering a call is the easy half. The value is in what happens next.

Our deployments earn their keep when the receptionist writes into the actual system of record. Booking into the real calendar with the correct appointment duration for that treatment type. Checking whether the caller is an existing patient. Tagging urgency and escalating to a mobile. Pushing a structured summary into the CRM so the follow-up happens without anyone remembering to do it.

Bundled receptionists handle the generic version of that well and the specific version poorly, because the specific version is your business logic and no packaged product ships with it. The gap shows up as a caller who was pleasantly answered and then booked into the wrong 20-minute slot. If you want the full comparison of renting versus building this, we wrote the build-vs-rent math for an AI phone receptionist with both paths costed out.

What we would do with a 100-minute plan

Use it as a diagnostic, not a solution.

Run Zoom's free trial for two normal weeks on your live number. You get three things out of it that are worth more than the $29.99: your real inbound minute count, your real average call length, and a list of the exact questions callers ask that the generic script fumbled. That list is the specification for whatever you build or buy next.

If the trial shows you comfortably under 100 minutes, stop there and keep the $29.99 plan. That is a good outcome and you should take it.

If it shows 400 minutes and a booking flow the bundled agent can't complete, you now have the numbers to justify a proper build, and per-minute platform pricing plus real integrations will beat the bundle on both cost and outcome. That is the AI receptionist we build and run for clients, on the number you already have.

And if the honest answer is that you only miss calls after hours and nobody needs to be booked live, the cheapest fix is not an AI receptionist at all. It is missed-call text-back, which we compared head to head against a full receptionist. Sometimes the $30 product is still $30 too much.