Data classification with AI
A model codes October’s journal lines to the chart of accounts, each with its confidence. Sure answers fill themselves in, the rest wait for one key or a person’s call.
6 stepsWatch an accountant code October’s journal: import the lines, let the model suggest every account, take the sure ones, decide the line it could not read, and post. Click into the grid at any point to take over.
- Import
- Account suggestions
- Review
- Post
Model run
Waiting for the ledger- Filled in by the model
- 0≥ 90% · 0 of 0 as booked
- To accept
- 0one key each
- To review
- 0under 40%
- By you
- 0taken or corrected
Suggested accounts by class
- 1 Non-current assets0 Dr 0 · Cr 0
- 2 Current assets0 Dr 0 · Cr 0
- 3 Liabilities0 Dr 0 · Cr 0
- 4 Equity0 Dr 0 · Cr 0
- 5 Revenue0 Dr 0 · Cr 0
- 6 Cost of sales0 Dr 0 · Cr 0
- 7 Administrative expenses0 Dr 0 · Cr 0
- 8 Other income and expenses0 Dr 0 · Cr 0
- 9 Income tax0 Dr 0 · Cr 0
Skip past the gridSkip back before the grid
Import October’s exports
One export per document type, from five group entities. For every line the model proposes a debit and a credit account from the group chart.
- supplier-invoices_2026-10.csvSupplier invoices · 105 lines
- customer-invoices_2026-10.csvCustomer invoices · 22 lines
- bank-statements_2026-10.csvBank statements · 36 lines
- payroll_2026-09.csvPayroll · 30 lines
- expense-claims_2026-10.csvExpense claims · 6 lines
- petty-cash_2026-10.csvPetty cash · 1 lines
Or drop any file here: the demo imports its own sample.
Make every data point count. Turn millions of them into actionable wisdom.
Let’s make it easier.
Smoother. Faster. All Handsontable.