A 3v3.ai capability · the studio where Yi.X products are born

The engine room of the Yi. Suite.

Every product on this site started as a bespoke build for one customer. Someone walked in with a problem nobody else's software solved. We built it. Then we noticed - this works for more than one. So we productized it. Six products. One pipeline. Yours could be next.

6
products born here
488
clients in production
5+
live integrations
1
accounting firm
The story we usually tell second

We're an accounting firm that ships software.

Six products live on the rest of this site - you can click through to each one. Yi.zy. Yi.ncome. Yi.nvoice. Yves.Lens. Yi.ngdom. Yi.Flow. They look like a typical software family. They're not.

Every one of them started the same way: a real customer walked in with a real problem, nobody else's software solved it, and we built the thing that did. Then we kept looking at the result and thinking, wait - this works for more than one of us. So we productized it. Made it a SKU. Shipped it to anyone else who needed it.

That's not a marketing story. That's the operating model. The Yi. Suite is the residue of an accounting firm solving its own problems and its customers' problems, then realising the solutions were re-sellable. Every product is proof that this pipeline works.

AI Plumbing is the studio where that happens. The build bench. The pipeline before the SKU. It's the first step of every product we already ship - and the first step of every product we'll ship next.

The Yi. Suite is what came out.
AI Plumbing is what goes in.

The bench

The pipeline didn't start with AI.

Long before bespoke builds became fashionable - and long before AI made them affordable - our accountants were doing the same work with the tools of the time. The six SKUs on this site are the AI-era chapter of a much older practice.

Era 1 · The desktop years

"Profit", customised - one client at a time.

Our firm built Profit, a desktop accounting application. On the surface every install looked the same. Underneath, each one was tailored to the client we'd handed it to.

Industry build

Gyms

Membership accounting and class-revenue recognition wired into the GL.

Industry build

Transportation

Fleet-cost allocation, trip-level P&L, route-by-route profitability.

Industry build

Construction

Project accounting and progress billing through the full project lifecycle.

Treasury build

5 banks · 3 currencies

Multi-bank reconciliation and consolidated FX position - the way a real treasurer thinks.

Reporting build

SGD presented · USD functional

IAS 21 dual-currency reporting. IRAS-clean on the SGD side, true economics on the USD side.

Bespoke build

Wrestling Federation of Singapore

Non-profit accounting with grant reporting - because off-the-shelf software doesn't ship for federations.

Era 2 · Post-desktop, pre-AI

The work didn't pause when "no-code" got popular.

When the desktop era ended, we kept building in the seams - with whatever tools the moment offered.

Automation

Google Apps Script

Sheets and Docs automations - the kind of glue that lives between the systems your accounting software refuses to talk to.

Internal app

AppSheet

Mobile and desktop database apps built on Google Workspace. The Sheets the firm already ran on - finally usable on a phone.

Integration

Zoho → OCBC batch upload

Custom batch-file generator that turned Zoho output into bank-ready uploads. A clerk's afternoon compressed to seconds.

Era 3 · The early AI years

We built our own MCP - before MCP existed.

When Anthropic shipped Model Context Protocol and "Claude.ai connects to Xero / MYOB" became a one-click pattern, we'd already shipped it ourselves. Our CEO was asking Yi.ngdom for dashboards on live Yi.ncome data - in natural language - before "MCP" was a term.

yingdom.3v3.ai / ai_dashboard
AI Dashboard Generator - natural-language financial dashboards built on live Yi.ncome data
Chat-to-dashboard

AI Dashboard Generator

Natural language → SQL → Yi.ncome data → rendered dashboard. "Show me sales revenue for Q1 2026" - and there it is, drawn from the firm's live books.

Same pattern Claude.ai's MCP shipped a year later. We just built it first, on our own data, for the CEOs who couldn't wait for the rest of the industry to catch up.

All of it written by the accountants in our firm - the same people who do your books.
At the bench, we don't have devs. We have accountants who code.
When a build outgrows the bench, the developers join - full-stack, AI, UI/UX, server - to scale, speed, secure, deploy.
AI didn't create this bench. It just gave us a faster keyboard.

The flywheel

One repeating motion.

Bespoke builds become products. Products fund the next bespoke build. Customers underwrite SKUs that benefit the customer after them. The model compounds.

01 · The signal

A real customer, a unique problem

Someone walks in with a constraint nobody else's tool solves. Often a Singapore-specific edge case. We listen.

02 · The build

We ship the bespoke solution

Discovery, wireframe, build, handover. Custom architecture, not a template. The customer gets exactly what they need.

03 · The check

Does this generalise?

If only one customer has this problem - they get a moat. If many customers have it - we productize. The same code becomes a Yi.X SKU.

04 · The SKU

It joins the family

New product on this site. Pricing tier. Marketing page. Next customer pays a fraction of what the first one did - and that customer's problem feeds the next bespoke.

The customer who underwrites the first build gets it built for cost, helps birth a product, and now has a moat that competitors will pay full price to license. That's the unfair model.

The proof

Six SKUs. One pipeline.

Every Yi.X product you can buy started here. Click through to see what each became - and read the rest of this site as evidence that the model works.

How we work

From your problem to shipped code, in four steps.

No slide decks. No 6-month "discovery phases". We diagnose in 30 minutes, wireframe in a week, ship in weeks - not quarters.

01

Diagnose

30-minute call. You tell us the problem and what's broken about every existing tool. We tell you whether we can build it - honestly.

30 min · free
02

Wireframe

Within a week we send back a mockup of the system and a scoped quote. You approve, redirect, or walk - no obligation.

~1 week · fixed scope
03

Build

Typically 2-8 weeks depending on integrations. You see progress every Friday. Production deploy on our infra or yours.

2-8 weeks · weekly demos
04

Generalize (optional)

If your build generalizes, you can choose: keep it as a moat (you own the code) or co-launch it as a Yi.X SKU (you share the upside).

Optional · revenue share
Why this beats the alternatives

Three other paths. One that ships.

Every business has the same four options when an off-the-shelf tool doesn't fit. Here's what each one actually delivers.

vs SaaS (Xero, QBO, etc.)

Built for everyone

They sell finished products. They can't bend to your problem - their roadmap is decided by 50,000 other customers, not you.

vs Consulting firms

They sell slides

The deliverable is a deck and a PowerPoint roadmap you implement yourself. We ship the code. The deck is included for free.

vs Hiring developers

They sell hours

Devs bill time. We bill outcomes. Six shipped products and 488 clients in production say we know which builds matter and which don't.

vs Building in-house

You learn what we learned

Building accounting + AI infra from scratch takes years. We already built the bones. Your build sits on shoulders - not greenfield.

In flight

What's coming next.

One product is always in the studio, halfway between bespoke and SKU. Here's what's brewing now.

In studio

AI Audit for SMEs - sampling reborn for the AI era.

For 50 years, audit has been sampling-based - the auditor checks 5-10% of transactions. AI doesn't sample. With Yi.zy + Yi.ncome as the data substrate, ET Management can offer 100% population audit at SME prices - bringing audit rigour back to companies that grew up without it. The substrate is live. The audit service is the next SKU.

You have a problem. We've built the engine for problems like yours.

A 30-minute conversation about what's not working, what you've tried, and whether the studio can ship you what off-the-shelf software can't. No slides, no NDA paperwork up front, no sales pitch.

Built in Singapore · Shipped from an accounting firm · Productized when it generalises