Rosters, stock, job costing, quoting, compliance registers. Some spreadsheets should stay spreadsheets. Here is how to tell which yours is.
Almost every Australian business has one. A spreadsheet that started as a quick way to track something and is now, without anyone deciding it, the system the business depends on. The roster. The stock file. The job costing model. The quoting calculator with the pricing rules built into it. The compliance register that tells you whose induction expires next month.
The decision to replace spreadsheets with custom software in Australia is usually made too late — after a version conflict costs a day of work, or after a formula has been quietly wrong for a quarter. It is a better decision to make deliberately, while the spreadsheet still works.
Be clear about the other half of this. Plenty of spreadsheets should stay spreadsheets — a spreadsheet is the fastest modelling tool most businesses have. If one person uses a file to think, to model a scenario or test a price change, leave it alone. Building software around exploratory work makes it slower and worse. The same goes for anything used a handful of times a year, or where the point is that you can restructure it on the spot. The spreadsheets worth replacing are the ones several people depend on, that run on a schedule, and that other work is queued behind.
We are Drawn AI, an Australian business started in 2024 by Liam and Jacob — one from retail operations management, one from mining operations. Both of us have been the person maintaining the file everyone else relies on. We build replacements that connect to what you already run: Xero, MYOB, Shopify, Cin7, Deputy, SimPRO, Employment Hero, SQL Server.
Rostering with award-relevant rules, availability, shift swaps and a mobile view for staff, feeding approved hours into payroll rather than into a second spreadsheet someone retypes each fortnight. This is common in multi-site retail.
Live stock levels across sites, transfers, adjustments and stocktake counts done on a phone in the aisle, reconciled against your POS or inventory system rather than against a file emailed on Friday.
Labour, materials, purchase orders and variations tied to a job number, with committed cost visible before the invoices arrive. On construction and trades work, this is where multi-tab costing workbooks most often fall over.
Customer-specific pricing, margin floors, approval thresholds and version history — so a quote cannot be sent below margin because someone copied last month's tab and missed a cell.
Inductions, licences, insurances, permits, SWMS and plant certificates, with reminders before expiry and a record of every change. Registers are the strongest case for moving off a spreadsheet, which cannot prove who changed what.
Once the data is structured, dashboards and reporting become straightforward. Most reporting pain is really a data-structure problem.
We look at the actual file, not a description of it. The logic that matters is usually buried in three columns and one person's habits, and has to be understood before anything is rebuilt.
We separate what the spreadsheet does from what it should do. Files that have grown over years carry rules nobody needs any more, and mapping the process properly keeps you from paying to rebuild those.
We build the replacement and run it alongside the spreadsheet for a period, with the same data going into both. Nothing gets switched off while it is still unproven. You find the differences before you rely on the new one.
We extract the history, clean it and load it — including the parts of a spreadsheet that fight back: merged cells, dates stored as text, blank rows carrying meaning, four spellings of the same supplier name.
Switch when the parallel run says it is safe, usually at a natural boundary like the start of a pay cycle, month or stocktake. Then we refine it, often adding automation to the steps still done by hand.
Three tests. Do several people need it at once? Does other work stop when one person is away? Would you be comfortable showing its history to an auditor or insurer? If the answer to any of those is no, the spreadsheet is doing a job it was not built for. If all three are fine, keep it.
Drawn AI works on a monthly subscription — minimum three months, billed monthly, 30 days' notice to cancel — rather than a fixed project price or an hourly rate. The figure depends on how much the spreadsheet does and what it needs to connect to. We give you a monthly number before you commit.
No. We build the replacement alongside the existing file and run both for a period with the same data going into each. You cut over at a natural boundary — end of a pay cycle, month or stocktake — once the parallel run shows the two agree. The original file stays as a reference.
It comes with you. We extract, clean and load the historical data, and we tell you plainly what is recoverable and what is not — old rows with missing fields, inconsistent naming, deleted tabs. Getting clean history in is often the most valuable part of the work, because it is what makes reporting worth anything.
Sometimes, and if a product fits we will say so. Configurable tools work well when your process resembles the one they were designed around. They get expensive and fragile when it does not, and you can end up rebuilding the same workarounds you had in the spreadsheet. That is the point at which custom software built for your process pays.
Yes. We are based in Melbourne, Victoria and Queensland, and work Australia-wide, remotely. Spreadsheet replacement work is well suited to it — the file, the data and the review sessions all travel. Send the file and a short description to outreach@drawnai.app and we will tell you what we would do with it.
A first conversation costs nothing and usually ends with a clearer idea of what is worth building — sometimes that answer is “not yet”, and we will say so.
No lock-in templates. Built around you.