You answer five questions about what it should do. We build it, you test it, and it goes live on your website, your phone line and your texts. No toolkit to assemble and no engineer to hire.
Fixed price, quoted before you pay. Live in 7 working days on Starter.
Live chat on your site
Do you have the navy runners in a 9?
Two pairs left in navy, size 9. I can hold one at the counter until 6pm tomorrow.
Held in Shopify
Inbound call, transcribed
Are you open Sunday? I need a cut.
We are, 10 to 4. Sunday at 11 is free. Shall I put you down for it?
Booked in Calendly
Text message
where is my order
It left the warehouse this morning and is due Thursday. Want the tracking link?
Read from your store
How it works
Ten minutes, in your own words. It asks follow-ups about what you actually said, not a form with forty fields.
What it will do, what it will connect to, what it will refuse. Edit anything that is wrong. The price is on the same screen, before you pay.
Six stages you can watch from your dashboard, each with a note from the person doing the work. You connect your accounts when we ask.
Talk to it in a sandbox first and tell us what to change. Nothing goes live until you press the button.
Pricing
Half the build price when you approve the spec, half when it goes live. The monthly plan starts the day it is live and covers hosting, monitoring and support.
An agent on your website that answers questions.
$1,500 build
$199/mo once it is live
Most common
It answers on every channel, and it can act.
$4,000 build
$499/mo once it is live
Inside your codebase, your deploy, or regulated data.
$12,000 and up
From $999/mo once it is live
What it runs on
An agent that touches real customers and real money needs more than a good prompt. This is the layer underneath every build, and it is the same layer on all three tiers.
Identity is bound server side, before the model sees a tool. It cannot be talked into reading somebody else's order.
Refunds, cancellations and deletions wait for a human. You set the limit above which it always asks.
Full transcripts, every tool call, every input and output. Flag a bad answer and it becomes a test case.
Idempotency keys on every write, so a dropped connection does not bill your customer twice.
Model spend is metered per conversation and capped per month. It stops rather than surprising you.
One switch pauses the agent on every channel at once, without taking your site down with it.
Try it
This is an AXONIC agent answering for AXONIC. It knows the pricing, the process and what we will not do.
The live demo needs JavaScript. Everything it would tell you is on this page, and the intake will ask you the rest.
Before you start
It only ever sees your own data, and anything that changes something waits for a person. When it does not know, it says so and hands off rather than guessing. You get the transcript for every conversation, so you can see what it was asked, what it read and what it said, and flag it.
You own the spec, the content, the conversations and the customer data. The platform it runs on is ours. You can export every transcript at any time, and closing your account deletes your data rather than archiving it.
Yes. Raise a change request from your dashboard. We come back with a price and a timeline, and nothing changes until you accept it.
You only pay the balance once you have tested it in the sandbox and approved it going live, so you are never on the hook for a build you have not seen work. After that the monthly plan runs like any other subscription. Cancel it and it stops.
Not for Starter or Pro. Connecting an account is a sign in, and we ask for the narrowest access that covers your spec. Bespoke is the tier where somebody technical on your side is usually involved, which is part of why it is quoted on a call.
You do not pay anything to find out what it would cost. Answer the questions, read the spec, and decide then.