Finance
Costs went up 18%. The entries say why.
AI reads the ledger, purchasing and payables in your ERP, follows the change from category to vendor to cost centre, and stops at the journal entries behind it. The report goes to the controller; sending it is still a person's call.
Where this starts
The quarter closes and operating expenses are higher than planned. The ledger has every entry, but the question from the CFO arrives as a sentence: why did costs go up?
What made it hard
Answering it means exporting the ledger, pivoting by account, then by vendor, then opening invoices one by one to see which ones are real spend and which are accruals. By the time it is done, the next close has started.
How it works
A semantic layer maps the ERP's ledger, purchasing and payables tables to the terms finance already uses — operating expense, vendor, cost centre, accrual. The question is resolved against it, each level shows how much of the change it explains, and the last step lands on the entries themselves.
Step through a run
Pick a question and follow it down — including one where the increase turns out not to be spend at all.
Pick a sample
Financial data is read from the ERP
Ledger, sales, cost and purchasing tables are pulled on a schedule. Nothing is copied by hand, and every figure below traces back to a row.
Read from
- ERP (SAP)fi.gl_line_item486,200 rows
- ERP (SAP)mm.purchase_order22,410 rows
- ERP (SAP)ap.vendor_invoice31,870 rows
- Period
- 2026-04-01 — 2026-06-30
- Last loaded
- Jul 1, 2026, 04:20
A recorded run against a sample ledger. The figures and vendors are synthetic. What the semantic layer resolves depends on how your own ERP is configured.
What it runs on
- ERP ledger, purchasing & payables
- Semantic layer
- Natural-language query
- Report generation
Spending days explaining variances?
Tell us which numbers you are asked to explain every month. We will tell you whether they can be answered from the ledger directly.
Talk to us