I shipped my AI receptionist on Product Hunt today — here's what I learned

Casey Smith — Tue, 09 Jun 2026 17:51:16 +0000

Today we launched Telyn(https://telyn.ai/ph) on Product Hunt — an AI receptionist for small businesses. It's $97/mo, answers in under 2 seconds, 24/7, and books appointments directly into Google Calendar mid-call.

Sharing what I learned shipping it, for any other indie founders pre-launch.

The problem we kept seeing

A friend who runs a 4-truck HVAC business showed me his call log one Saturday. 71 missed inbound calls in a single weekend. Every single one went to voicemail and never called back.

He'd hired and fired three "answering services" over two years. They missed bookings, couldn't access his calendar, and quit at 5 PM on Fridays. He told me: "all I need is something that picks up the phone and books the job."

The numbers on this problem are brutal across SMBs:

Average small business gets 30-120 inbound calls per week
40-60% of those calls come after business hours
Voicemail catch rate is ~4%
Which means 96% of after-hours callers hang up and never call back

For HVAC, dental, plumbing, real estate, legal, property management — anyone whose lead-gen depends on the phone — this is silently bleeding $20K-$80K/year per business.

What was already on the market (and why it didn't work for SMBs)

Before building, I tested every product in the category:

Human-staffed answering services ($300-600/mo): Smith.ai, Ruby. Beautiful voices, brand recognition, but neither could book directly into Google Calendar without a manual sync layer. And the pricing is out of reach for solo operators.
First-generation AI products ($200-400/mo): Decent voice but needed a developer to wire up booking. Most SMB owners don't have a developer.
Newer AI entrants: One required developer setup (kills it for SMBs), the other was free but voice quality was a deal-breaker.

The whole category was either too expensive, too technical, or too rough on voice. Nobody was building for the SMB owner who just wants the phone answered.

What we built

The core promise: pick up the phone in under 2 seconds, book the job, text the owner.

The implementation:

Voice agent layer with sub-second TTFB targeting
Direct Google Calendar integration for appointment booking mid-call (no Zapier required for the happy path)
SMS confirmation to the caller before they hang up
Full call transcript SMS'd to the business owner instantly
Smart routing for emergencies (frozen pipes, no-cooling in summer) → on-call cell phone
Bilingual EN/ES out of the box (critical for HVAC and home services in markets like Dallas, Miami, LA, Houston)
10-minute onboarding wizard — pick industry, customize call flow in plain English, forward your number, you're live

What we deliberately stripped out: flowchart builders, prompt engineering UI, multi-tier integration matrix. The first prototype had all of it. We watched our beta operators (HVAC, dental, law firm) get confused and bail out during setup. So we ripped 80% of it out.

The pricing decision

We landed at $97/mo Starter (500 voice minutes), $327/mo Growth (1,500 minutes), $697/mo Pro (2,500 minutes). 7-day free trial.

The math we cared about: $97/mo is 1/3 the price of the cheapest human-staffed service and easily outperformed two of them on call handling in our tests. We wanted the price to be unambiguously a no-brainer — not "comparable to" something but "obviously cheaper and obviously works."

We're locking in $97/mo for life for everyone who enrolls during the launch window. Not a promotional gimmick — it's a founder commitment that the price you sign up at is the price you keep, even when we raise list pricing in Q3.

What surprised us

The voice quality bar is way more important than the AI features bar. Beta users abandoned products with robotic voices in the first 30 seconds and never came back. We over-indexed on voice early. Worth it.
Calendar booking is the whole product. Everything else is table stakes. The first thing every operator asked: "does it actually put the appointment on my Google Calendar?" That's the entire value prop. Get that right, the rest follows.
SMBs care more about owner SMS than caller experience. The business owner getting an instant transcript on their phone matters more for retention than any caller-side feature we built. They want the receipt. The peace of mind.
Setup friction kills more deals than pricing. Our biggest churn cause in beta wasn't price — it was operators who started setup, hit a webhook config screen, and bailed. Stripping the wizard down to 10 minutes was the single biggest activation lift we shipped.

Try it like a customer

Public demo line: (580) 729-3227 — call it yourself, pretend you're a customer. Try to break it. Try to confuse it. Honest feedback in the comments would be appreciated.

PH launch: telyn.ai/ph

Happy to answer questions about the voice stack, latency targets, integration architecture, pricing logic, or what we're getting wrong. Drop them below.