How we work
Find it. Build it. Keep it running.
Most of our job is deciding where AI belongs in a business, and where it doesn't. We build the part that earns its place, into the systems you already run, and we stay to run it.
The idea underneath
Rules, AI, people
AI everywhere is bad design. Most of a business runs on fixed answers and on people, and it should.
Rules
For what is always the same: a price, what's included, whether a slot is open. Looked up from your systems, never guessed. Always right, cheap to run.
AI
Only where the input is messy: a customer types something half-remembered and it has to land on the right service.
People
Where judgement or accountability matters: anything clinical, anything that commits money, anything that needs a yes from someone who can be held to it.
Get the split right and there is less to maintain, less that can drift, and staff keep the tools they know.
01 · Find it
The first answer is often "not here".
We sit with the people doing the work and follow a few real jobs end to end. You get one page: where a rule would do, where a person should stay, and the one place AI would actually pay for itself. "Don't build this" is a good outcome, and we say it when it's true.
Sometimes the honest answer is a product, not a project. For a small appointment business that just needs bookings handled, we'll point you to FrontDesk instead.
Using a grant? Map first, and talk to us before signing anything. What the EDGE grant means if you want to use AI →
02 · Build it
Into what you already run.
Your WhatsApp number, your website, your booking or practice-management system. No new logins, no migration.
Written down first
The plan for what it will and won't do is written in plain words in the first week.
Fixed price
Quoted after the map, agreed before you commit.
Staged before it's live
Your team tests it on a staging copy before a customer sees it.
Reads freely, writes carefully
Anything that changes a booking, a record or an amount is guarded, logged and confirmed.
If you know the term, we work forward-deployed. We sit inside your team, learn how the work is really done, and stay accountable once it's live.
03 · Keep it running
Month six is what matters.
Prices move, services change, someone asks what nobody wrote down. An AI system announces none of it. It just starts being wrong, confidently. So every build comes with the loop:
1
Monitor
We check every conversation against what your business actually does. Alerts reach a named person with the conversation attached.
2
Pause
You can stop it yourself, at any hour. It also stands down on its own when it's out of its depth.
3
Correct
Your staff correct a wrong reply in plain words and it takes effect the same day. No developer, no redeploy.
4
Test
Every correction becomes a test that must pass before anything ships.
The question to ask any vendor, including us: when it tells a customer something wrong, how will you find out?
One live deployment, four locations, June to August 2026: 90 corrections worked through, 37 → 99 tests guarding them, none deployed without review.
Working with agencies
Your clients are asking about AI. You don't need to build an AI team to say yes.
Bring us in on a client's project
We work as part of your delivery team, under your name or ours. You keep the relationship; we map it, build it and stay on to run it.
Or refer them
Introduce a client, and we take it from there and keep you in the loop.
An NDA. We never approach your client for other work. Everything is staged before it goes live. Commercial terms agreed with you per project.
How pricing works
No surprises, in this order
- First, a conversation about where AI fits in your business.
- S$1 for each booking made through the assistant. No monthly fee.
- Anything built into your own systems depends on requirements and complexity. Speak to us: it's quoted up front and agreed before you commit.
- Grants: if you're applying, you apply yourself. Talk to us before signing anything, so the order is right.
Tell us what you're trying to build
For your own business or a client's. We'll tell you honestly whether AI is worth it, and where we'd start.