11 years ago, one accountant refused
to enter the same number three times.
That stubbornness became a system. That system became MyLedger.
This is the story of how a late-night spreadsheet experiment turned into an AI-powered accounting platform — and how an entire industry arrived at the same conclusion we reached a decade ago.
Read the full founder's letterThe Midnight Problem
Sometime in 2015, well past midnight
The office was empty. The coffee was cold. And spread across the screen was the same thing as yesterday — hundreds of lines of bank transactions from a client who kept no books, used no software, and had no idea what half those charges even were.
A T-account. The kind every junior accountant in every firm in the country has been trained on since day one. Left column: description. Right column: amount. Then a separate column for classification. One transaction, manually keyed three times. Get one cell wrong, and the rest of the sheet collapses.
GST had to be hand-calculated. Categories had to be hand-assigned. Every row was a small, tedious act of faith — faith that you hadn't already mistyped something four rows back.
This wasn't a broken process. This was the process. The whole industry had accepted it as the cost of doing business.
“Why does this have to be so painful?”
That question, asked in the dark of an empty office, wouldn't go away. And it changed everything.
11 Years of Innovation
From a late-night spreadsheet experiment to a full AI-powered accounting platform.
The Breaking Point
Our founder was drowning in messy client bank statements every day, working late into the night. Every transaction had to be entered three times in old-fashioned T-account Excel sheets — once for the entry, once for the amount, once for the category. Every junior accountant in Australia has done this. It was slow, painful, and error-prone.
The Eureka Moment
While searching for a better way, our founder discovered that flat-data management — like an Excel Pivot Table — could dramatically speed up transaction categorisation. Instead of T-accounts, he laid out bank data in a flat table: Date, Narrative, and Amount. Three columns, pulled straight from bank statements.
The Five-Column Foundation
Two more columns were added: a Comments column for clients to describe each transaction's nature, and an Accounting Category column mapping to the Chart of Accounts. With these five columns — Date, Narrative, Amount, Comments, and Category — Pivot Tables could generate any report needed. This was the birth of what we internally called "Tax7Data" — written as one word, no space. Why? Because some accountants kept typing "Tax7 Data" with a space, which broke our early VBA macros. We couldn't auto-strip spaces reliably, so we made it a rule: Tax7Data, always.
Expanding the Data Model
As client needs grew, we added Business Percentage for mixed-use expenses, GST Coding for tax classification, Income Tax Return Labels for ATO reporting, and Financial Ledger Labels for generating financial statements. Every line of automation in those early years was written in Excel VBA — coded late at night after a full day of client work. Each column solved a real compliance requirement through endless trial and error. Nothing was theoretical; everything was born from real pain.
The HandiLedger Breakthrough
Tax7Data connected to HandiLedger — and for the first time, we achieved full automation from Bank Statement all the way to Financial Statement. Tax7Data could feed directly from a local machine into HandiLedger and generate financial reports with a single click. At the time, this was a major breakthrough: an end-to-end pipeline that no one else had.
Proven at Scale
Tax7Data became the engine behind Tax7, a multi-million dollar Australian accounting firm. But scaling meant outgrowing Excel VBA entirely — we rewrote years of battle-tested logic into a modern codebase, preserving every lesson learned. Over 1 million transactions were auto-reconciled with minimal staff. The flat-data concept that started on a single spreadsheet was now powering an entire practice.
The Xero Meeting
Our founder met with Xero's product management team, seeking to connect Tax7Data with Xero's Bank Feed to further automate the workflow. During the meeting, he demonstrated the flat-data reconciliation approach in detail. The partnership never went ahead — but the experience reinforced our conviction that flat-data reconciliation was the right direction. When Xero later introduced Cash Coding, a spreadsheet-style reconciliation interface, it felt like validation that the industry was moving toward the same conclusion we'd reached years earlier.
The QuickBooks Visit
Post-COVID, our founder — still wearing a mask — visited QuickBooks in person to explore importing Tax7Data clients into their system. After considerable effort, it became clear that QuickBooks' database architecture simply couldn't accommodate our flat-data transaction format. The integration was abandoned. It was becoming obvious: to do this right, we needed our own platform.
Going Independent
After years of trying to fit Tax7Data into other people's software, our founder made the decision: rebuild it from the ground up as a standalone web application. No more dependencies on third-party platforms. A massive investment of time, energy, and resources — but the only way to truly deliver the vision.
AI Meets a Decade of Data
When AI capabilities matured, we had something no competitor could replicate: 10 years of structured, real-world accounting data and battle-tested workflows. We integrated AI-powered auto-reconciliation, intelligent categorisation, and automated working paper generation — all built on the same flat-data foundation.
MyLedger is Born
Tax7Data evolved into MyLedger — a full platform for accountants with auto-reconciliation, BAS/ITR reporting, AI Copilot, and working paper automation. The flat-data philosophy never changed. What started as one accountant's late-night experiment is now ready for every accounting practice in Australia.
Tax7Data: The Flat-Data Revolution
One unified data format that replaced T-accounts, eliminated triple-entry, and enabled automated reporting. Each column was added to solve a real compliance need — nothing theoretical, nothing wasted.
Transaction date from bank statement
Bank description of the transaction
Debit or credit amount
Client explains what the transaction is for
Chart of Accounts mapping
Percentage used for business purposes
GST classification for BAS reporting
Income Tax Return label for ATO
Financial statement classification
= Pivot Tables, BAS, ITR, Financial Statements, Working Papers — all from one flat dataset
What It Took to Get Here
What you see today is the result of 11 years of trial, error, late nights, rewrites, and relentless refinement. None of this came easy.
Years in Excel VBA
Every early automation was hand-coded in Excel VBA, written late at night after full days of client work. The first versions were rough — but they worked.
A Complete Rewrite
When we outgrew Excel, we had to rewrite years of proven logic into a modern stack — preserving every lesson while rebuilding from scratch. That took years more.
1M+ Transactions
This flat-data approach was tested and refined through over a million real transactions — not in a lab, but in a live accounting practice serving real clients every day.
Who We Built This For
MyLedger was never built to impress investors or chase trends. It was built for a very specific kind of client — the kind that most systems quietly exclude.
The small business owner who can't afford a bookkeeper.
The sole trader who tried Xero once and gave up after a week.
The client who shows up in March with a shoebox of receipts and a bank statement full of mysteries.
The one whose records are so messy that most accountants would rather decline the engagement than untangle it.
In Australia, there are hundreds of thousands of small businesses and individuals who fall behind on compliance — not because they don't care, but because the tools that exist weren't built for them. The software assumes you already know what you're doing. The pricing assumes you have budget for professional bookkeeping.
MyLedger assumes none of that. It assumes your data is messy. It assumes you're behind. And then it gives your accountant the structure to turn that chaos into a compliant outcome — a lodged tax return, a filed BAS, a set of financials that actually reflect what happened.
One data structure. Eleven years.
From chaos to compliance.
This is MyLedger.
Experience 11 Years of Innovation
See for yourself why our approach to reconciliation is being adopted across the industry. Get started with MyLedger today.