What happens the first time the AI says something stupid to someone about to pay me $8,000?

Nobody can promise it never will, and a vendor who promises that is managing your perception rather than your risk. Two things make it survivable. It says what it is at the start of the call, so no registrant finds out afterwards that they were handled under false pretences. And the first run is a split of your own list, so the blast radius of a bad call is half of one webinar.

Does it tell people it's an AI, and what if someone asks directly?

It identifies itself as an assistant at the start of the call, before anything else — so no registrant ever discovers something was hidden. The word-for-word line carries legal weight, so it lives on the what it says page and gets read to you on the call.

Can I hear a call that went badly?

Not on this site today, and we would rather say so than post a reel of the good ones. When publishable calls exist, one that went badly ships next to them, labelled. The hear real calls page explains what has to be true first.

Not published here, and that is a policy rather than an evasion. The shape is a base plus a fee per attended call, with a pilot before any of it. The number gets named on the call — straight away, before anyone asks what you spend on ads.

Show-up rate is a vanity metric — did anyone actually buy, and what's your worst client outcome?

There have been no clients, so there is no client outcome, good or bad, and nothing on this site is a case study, a logo or a testimonial. There is also no traceable benchmark anywhere for how many no-shows eventually buy — the figures in circulation trace back to vendor pages with no method attached. The only answer worth anything is measured on your own list against your own untouched half, on revenue per registrant.

For webinar operators

You run a webinar. We increase your show-up rate and your post-webinar sales, with our AI voice and SMS system.

We text every registrant the moment they sign up and ask why. Then we call before the session and get them to commit to a time
— and you keep every answer.

Book a call

No deck. No demo. The first run is a split of your own registrant list.

It says it's an assistant.

You approve every line.

Half your list is the control.

Zero clients. We say so.

phases — one we do not sell: 4
your list, untouched, as the control: 1/2
line at the top: it is an assistant: 1
clients closed. no case studies.: 0

The gap

Registration and people who show up
are completely different topics.

Your ads buy registrations. Your offer converts the people who turn up. The number in between is your cost per booked-and-attended call. Fill it in for one webinar.

Worksheet

Ad spend behind one webinar: [yours]
Registrations it bought: [yours]
Cost per registration: [1 ÷ 2]
Of those, how many attend live: [your rate]
Cost per booked-and-attended call: [3 ÷ 4]
The only line on this list we can move: line 4

The closest thing to a benchmark

21%

238 registrations, 51 live attendees — the average webinar in Goldcast's 2025 dataset. It headlines 33%; its own averages imply 21%, which is the one you can recompute.

Run it on your numbers

No published benchmark segments show-up rate by traffic source, and the largest disclosed dataset measures marketers emailing their own list. What a normal show-up rate looks like.

Why doesn't another reminder fix it?

people in the randomised field experiment: 287,228
movement from a standard encouragement call: 0
when the call got them to state their own plan — among the people it reached: +4.1pp

That is a micro-commitment — a sentence they said out loud, not a fifth notification. Whether it holds when the caller says it is an assistant is what your first split test settles.

Nickerson & Rogers, Psychological Science 2010, N = 287,228 · +4.1pp treated, +0.9pp intent-to-treat · Why registrants don't show up · What actually raises attendance

The second product

What do you get besides a higher show-up rate?

An email sequence knows names and open rates. This ends with you knowing why every registrant signed up, in their own words, every week. Nobody else asks them anything, so nobody else can hand you that.

One question, four places it pays
When	What the answer is worth
On the day	You know what people signed up to solve, not a headcount and a list of first names.
After the pitch	Your follow-up answers a concern somebody actually raised, in their words.
In the offer	Verbatim customer wording — you cannot write that from an open rate.
In the ad account	Stated motivations feed back into targeting.

The four phases

What actually happens to one registrant?

Four phases, and we name the one we do not sell. The touch-by-touch schedule gets named on the call, not estimated by a web page. The sequence, in full.

The four phases
Phase	Trigger	What happens
01 Instant opt-in	Registration submitted	A two-way text that says it is an assistant, then asks why they signed up.
02 Pre-webinar	15 minutes to 2 hours before	A call and a text that use their answer and ask them to say when they will join. That is the micro-commitment.
03 Mid-session exit	Attendee leaves before the pitch	The part we do not sell.
04 Deadline blitz	Post-event until the offer deadline	Calls, texts and email answering the concern each person stated in phase 1.
Why phase 03 is not sold: Zoom is the only platform with a genuine live event for it, and Zoom's own developer forum carries an unresolved report of that event silently not firing.

The deliverables

What do you actually get?

Four things, not a feature list. The wording it uses is published word for word, and what the calls sound like has its own page.

A two-way thread with every registrant
A call before the session that asks for a plan
A weekly document of your registrants' own words
Compliance handled in the open — the responsible party named at the start of every artificial-voice call, calling windows in the recipient's local time, revocation honoured, and consent captured in your name, never pooled across clients.

The limits

Who is this not for?

Every vendor in this category claims everything. Here is where it does not work. If your registrants are simply waiting for the recording, that is a different problem.

This is for you if

You run live webinars and collect a phone number at registration
Enough registrants per webinar that splitting the list in half means something
You will read and sign off the script before anything sends
You sell something after the webinar — this is judged on sales, not attendance

This is NOT for you if

You want software to log into. This is a service, not a platform.
Your registration form has no phone field, or its consent language never mentions AI or an artificial voice
You want people called the second they walk out mid-session. We do not sell that.
You want a case study first. There are zero closed clients and we are not going to invent one.

The first engagement

Split your registrant list down the middle.

Same ads, same week, same offer. Half your registrants get the calls and texts, half get whatever you do today — judged on what you sold, not attendance alone.

01
You approve the script — and keep the stop
Nothing sends until you have read the opening line, the questions and the subjects it refuses to discuss. One word from you stops it mid-campaign.
02
We split the list 50/50
Not a thin test slice — an even split reaches a real answer on far fewer registrants.
03
One webinar runs both ways
Your untouched half is the only honest comparison you will get, and it costs nothing to keep.
04
You read the result and the words
Both halves measured on revenue per registrant, plus every reason your registrants gave.

Month to month
A kill switch on your side
You approve every line
A complaint pauses the system

Book the pilot call How the pilot works

No case studies here, because there are no clients yet. Pricing is not on this site — the shape is a base fee plus a fee per attended call, named on the call. Why it is not on the page.

The questions we get

Honestly answered.

Twenty minutes

Pick a slot.

We name the price on the call, before anyone asks what you spend on ads. Bring one webinar's numbers.

No application gate, no deck, no demo — ad spend, registrations, and how many turned up live.

Book my call

You run a webinar. We increase your show-up rate and your post-webinar sales, with our AI voice and SMS system.

CallHush at a glance

Registration and people who show upare completely different topics.