Answers first
A reply inside a minute, any hour. Speed is usually what decides who gets the job, and it is the one thing a small team cannot do reliably by hand.
AI that answers, books and follows up.
Most enquiries go to whoever replies first. An intake agent answers in about a minute — at 9pm, on a Sunday, while you are up a ladder — asks the qualifying questions, and offers a slot from your real calendar.
What it actually does
Asks the right questions
How many rooms, which boiler, what date, is there parking. The details you would have to ring back for, gathered before you ever pick up the phone.
Books from your real calendar
It offers slots that genuinely exist and holds them. No double bookings, no "let me check and get back to you".
Chases the quiet ones
A polite follow-up on the enquiries that went cold. Nobody does this consistently by hand, and it is often where the easiest work is hiding.
Knows when to stop
Anything unusual, sensitive or expensive gets handed to you with the conversation so far attached. It is built to escalate rather than to guess.
Leaves a trail
Every conversation logged where you can read it, so you can see what it said and correct the pattern rather than one message.
How we get there
-
01
Audit
A fortnight watching how enquiries actually arrive and what happens to them. You get a ranked list of what is worth automating and what is not — that list is useful whether or not you build anything.
-
02
Prove one
We build the highest-value one for real and measure it against how you work now. If it does not beat the current process, we tell you.
-
03
Ship it properly
The tested version, wired into the tools your team already uses, with the guardrails and approval steps agreed with you.
-
04
Hand over
Your team owns it, with training and a clear view of what it is doing, what it is costing and how to turn it off.
Questions we get asked
Will it sound like a robot to my customers?
That is the main thing we work on. The agent is written in your tone, kept short, and told to hand over to a person the moment a question goes beyond what it knows. A bad AI reply costs you more than a slow human one.
Does it reply without me seeing it?
Only if you want it to. Most clients start with every message drafted for approval, then relax that once they trust it on the routine questions. You choose where the line sits, and you can move it.
What if it gets something wrong?
It will, eventually — so it is built to fail safely. It is given a narrow job, told what it does not know, and set to escalate rather than guess. Anything involving a price, a promise or a date can be set to need a human first.
Do we need to change the tools we use?
No. The point is to wire into the inbox, calendar and systems your team already opens. If it needs everyone to learn a new app, it will not get used.
How do you know whether it is working?
We agree the measure before building — usually response time and how many enquiries turn into booked work. If the numbers do not move, we say so rather than pointing at how clever the technology is.
Is our data used to train someone else’s model?
Not on the setups we build. We use business API tiers where your content is not used for training, and we will tell you exactly which provider handles what before anything is connected.
Which job would you hand over first?
Tell us where enquiries come in and what happens to them now. We will tell you which part is worth automating — and which parts should stay human.