Every team has work that keeps coming back and quietly eats the hours meant for the work that matters. You already know which part that is for you.
01Quotations built line by line, job after job
02Quantities counted off drawings by hand
03Month-end reports stitched together from five spreadsheets
ST—02Awakening
Put your own numbers on it.
How often does it happen? How long does it take? How many people touch it? Three numbers tell you whether it's worth changing — and sometimes it isn't.
We sit with your team, map every step of one workflow and find where the hours go. You get a written map and a clear first target before any code is written.
01Watch the work
02Map every step
03Pick the first workflow
ST—04Emergence
Then it all takes shape.
The steps we mapped come together as one system that works alongside your team. It reads every document, cross-checks every figure, flags what doesn't add up and drafts the first version. Your people review, decide and sign off.
01Works inside the tools you already use
02Tested on your real work before it goes live
03Starts with one workflow, grows only when it pays off
ST—05Cilia · SIL-ee-uh · noun
Thousands of tiny parts, moving as one.
Cilia are tiny hairs on a cell that beat together to keep things moving. We bring that same coordination to the everyday work inside your business.
Tell us how the work moves today and where it stalls. We'll say plainly whether AI agents are the right answer, and if they aren't, we'll tell you that too.
Fig. 5
Weeks
024681012+
Discovery call
Diagnose
Build
Operate
01
Diagnose
1–2 weeks
We sit inside your workflow and map where your team's judgement is spent, and where it's wasted. You receive a written brief: the three places AI agents would change your business most, what each would take on, and a fixed price to build them.
Fig. 6
Weeks
024681012+
Discovery call
Diagnose
Build
Operate
02
Build
3–6 weeks
We design agents around your real documents, systems and rules, not a demo. They read, reason and draft the work end to end. You'll work with a live version from week two and shape it with us each week.
Fig. 7
Weeks
024681012+
Discovery call
Diagnose
Build
Operate
03
Operate
Ongoing
Agents get better with use. We watch every run, retune them as your work evolves, and report monthly on what they handled and the hours they gave back.
Fig. 8
Your expertise sets the direction. We keep the work around it moving.
Cilia is an AI studio in Singapore. We build AI software around the way people actually work. Start with a specific task, make the routine easier and leave judgment with the people who know the business. The aim is simple: less busywork, more room for work that matters.
Questions
Will it replace my team?
It changes what your team spends its day on. Agents take the first pass: reading drawings, drafting quantities, preparing quotes, at the speed of software. Your people keep the judgement and sign off on what goes out.
Where does my data go?
Your work runs on Claude, Anthropic's AI, through its business API. Your data is never used to train models and is deleted from Anthropic's systems within 30 days. We keep only what your agents need, delete on request, and work under Singapore's PDPA.
What if it gets something wrong?
Every agent shows its reasoning. Each figure traces back to the source it came from, so a review takes minutes, not hours. Nothing leaves without your approval, and as your work changes, we retune.
Contact
Tell us where the work stalls.
Start with a 30-minute discovery call, no charge. We'll say plainly whether AI agents are the right answer, and if they aren't, we'll tell you that too.
