Proof of Concept · Fixed scope
A working prototype in 2 weeks, for $5,000. Then decide.
Validate whether your idea is actually buildable before paying for a full product. AI generates the predictable 70%; our engineers spend the time on the risky 30%.
Common situations
"We do not know if this is buildable. We do not want to spend $50,000 to find out."
A PoC is for the moment before you commit to a full product — when the technical question is still open and you need an honest answer, fast.
"Is this technically feasible at all?" — and no internal engineer can confidently answer.
Investors and bosses ask the question. You need something demonstrable, not a slide.A vendor quoted $50,000 and 3 months just to write the requirements document.
You will not commit to that without proof the idea works. The estimate itself is the risk.You tried building it yourself with Cursor or Claude — and got stuck on the hard parts.
The 80% AI generates is easy. The 20% it cannot is exactly where engineering pays off.Before → After
What changes in two weeks.
A PoC is a transformation, not a deliverable. Below is the honest version of what you walk in with on Day 0 — and what you walk out with on Day 10.
- "Is this even buildable?" — and no one in-house can answer with confidence.
- Nothing demonstrable to take to investors, leadership, or pilot customers.
- Vendor quotes are $50,000 and three months. The estimate itself is the risk.
- No basis for a go / no-go decision on a full build.
- A working prototype — hosted — that demonstrates the riskiest piece end-to-end.
- Source code in your GitHub organisation, yours from the first commit.
- Written validation report: feasible, conditionally feasible, or not feasible.
- Concrete cost, time, and risk estimates for the full build — or a clear no-go.
How a PoC differs from a build
The point is the answer, not the product. The prototype is how you get to the answer.
A PoC is built to falsify a hypothesis in two weeks. That changes how it is scoped, written, and reviewed.
Attack the riskiest piece on Day 2
If something is going to break the idea, it surfaces in week one — not at the end. Fail fast is built into the schedule, not a wall slogan.
70% AI-generated, 30% hand-crafted
Boilerplate, scaffolds, mock data — AI handles them in hours. Engineers spend their two weeks on the genuinely uncertain technical work.
You leave with a decision, not a debate
The deliverable is a written validation report: feasible, conditionally feasible, or no. Sample code, cost estimate, risk register, all attached.
How it works · Day 1 → Day 10
Ten business days. Demo on Day 4 so the direction can change before it is too late.
All progress is visible in GitHub in real time. Daily updates in Slack. No "wait until the end" delivery — the demo on Day 10 is just the formal version of what you have already seen.
Kickoff & hypothesis
60-minute kickoff. Narrow validation hypotheses to three or fewer. Define what "success" and "failure" each look like, in writing.
PM + engineer + customer · 60 min
Hit the biggest risk first
Start with the most uncertain piece of technology — the part most likely to kill the idea. AI generates the prototype skeleton; engineers go straight at the risky surface.
Fail-fast. Worst case is finding out in 48 hours.
Mid-report
PM walks you through what is working, what is not, and whether to continue, pivot, or stop. This is the safety valve — most courses correct here, not at the end.
30-minute call · written notes
Build the prototype
AI generates ~80% of screens, API surface, and seed data. Engineers complete the integration and the genuinely tricky logic. Slack updates every working day.
Daily Slack updates · GitHub real-time
Internal review
PM and engineer do an internal demo. Validation report draft is generated by AI (~80%), reviewed by humans, and prepared for delivery.
Catch-and-fix before customer demo
Final delivery
30–60 minute customer demo. Validation report, source code, and a next-steps proposal handed over. You decide what happens next — including not doing anything.
All artefacts already in your GitHub
Why this works at this price
The numbers behind the $5,000.
We can sell a two-week PoC for $5,000 because we have spent the last two years industrialising the workflow on our own SaaS products. The economics work; we are not subsidising it.
A typical custom-development quote for the same scope is $50,000. A failed PoC costs you $5,000 — the price of an honest answer.
Scaffolding, mock data, screens, and API surface come out of AI in hours. Engineers spend the two weeks on the actually-risky 30%.
Tsunagu (community marketplace), Dr. Unco (AI health checkup), and L-career (Labuan job platform) — live products we operate ourselves. Fortress, our next one, is in active development.
If the PoC validates, the $5,000 comes off the MVP invoice.
The PoC is designed to be cheap if the answer is "stop" and even cheaper if the answer is "go." Continue into MVP Development and the $5,000 you already paid is deducted, in full, from the next invoice. No vouchers, no expiry, no fine print.
What you can rely on
Five things we will not move on, in any engagement.
These are not stretch goals or marketing claims — they are operational rules that govern every PoC we ship.
Day-1 ownership
Source code and intellectual property are yours from the first commit. The repository is created inside your GitHub organisation, not ours.
NDA before tech
NDA and DPA are signed before any technical discussion — including the free initial assessment. Idea-only conversations are also under NDA.
Daily transparency
Slack updates every working day. Real-time visibility in GitHub. Day 4 mid-report. No "wait two weeks and see" delivery model.
$5K cap on failure
If the hypothesis is invalidated, you owe $5,000 and you have a written report explaining why. There are no overage charges. Failed PoCs are part of the contract.
Honest no-go
If on Day 4 we believe the idea will not work and additional weeks cannot save it, we will say so. Continuing to bill is not in our long-term interest or yours.
FAQ
Five questions before you commit $5,000.
Specific case not here? Write it in the message field on the form and we will reply within one business day.
What if the PoC fails?
Does the $5,000 really roll into the MVP?
Do you really finish in 10 business days?
Who owns the code?
What if our scope is unclear?
Start a PoC
Tell us the idea. Get a free initial assessment.
We will respond within 1 business day with a free AI-assisted initial assessment of feasibility, before you commit to the $5,000.
- NDA signed before any technical discussion — including the free assessment.
- Your call is with the engineer who would actually scope the PoC.
- If we think a PoC is the wrong fit, we will say so.
- $5K · 2 weeks
- 100% IP
- GMT+8
- Credit toward MVP
Request a free initial assessment
Required fields marked *