Use cases by function6 use cases

Finance automation with a controller's instincts built in

Finance is where automation pays back most reliably and where mistakes cost the most, so the design differs from support or sales. Extraction, matching and chasing are automated. Anything that moves money, changes a supplier's bank details or posts to the ledger passes through rules in code and a person's approval. The model reads and proposes. It does not decide.

The appetite is there. Ardent Partners reported in August 2026 that 58% of accounts payable teams are using or piloting AI, and that slow approvals and high exception rates are their two biggest problems, at 48% each. Fraud is the other reason to build carefully: the FBI's 2025 internet crime report puts business email compromise losses at $3.05 billion, and the classic attack is a polite email announcing new bank details.

In Europe, e-invoicing mandates have changed the job. Belgium's Peppol mandate started in January 2026, Poland's KSeF in February, France began receiving from 1 September 2026 and Germany's issuing obligation starts in 2027. Structured invoices remove the scanning problem and replace it with validation, matching and legacy systems that cannot produce the format. The pages here cover supplier invoice processing, receiving e-invoices, issuing XRechnung from an old ERP and chasing clients for documents.

Finance and accounting systems I can build

The controls that come standard

  • Every posting can be traced

    Source document, extracted fields, the rules that passed or failed, who approved and when. An auditor should be able to follow any entry back to the email it arrived in.

  • Bank details are verified out of band

    A changed IBAN is never accepted from the email that announced it. It triggers a check through a contact already on file, and payment waits until a person confirms.

  • Nothing posts twice

    Every write is keyed so that a retry, a duplicate email or the same invoice arriving as PDF and e-invoice cannot create a second entry.

  • The exception queue is the product

    Straight-through processing is the easy part. What makes finance teams trust a system is a review screen that shows the document, the extracted value and the reason it was flagged, side by side.

  • Much of it is not AI

    Structured e-invoices, three-way matching and billing calculations are ordinary code. The model is used where text is messy: free-text invoice lines, emails, scans and coding proposals.

  • Your ledger stays your ledger

    Xero, QuickBooks, DATEV, NetSuite or SAP Business One remain the source of truth. The automation feeds them and reads from them; it does not become a second set of books.

Frequently asked questions

Can AI do bookkeeping?

AI can do much of the preparation: reading invoices and receipts, proposing account and tax codes with a reason, matching payments, and asking clients about unclear transactions. It should not post or pay without a person, and the accountant still owns the books. The realistic gain is that people review and decide instead of typing and chasing.

How do I automate accounts payable?

Start with capture and validation from the AP inbox, add matching against purchase orders and goods receipts, then approval routing in the tools your approvers already use, then posting. Bank detail changes and anything unmatched go to a person. Most teams see value from capture and matching before posting is automated at all.

What do the EU e-invoicing mandates mean for my business?

It depends on the country. Germany has required businesses to be able to receive structured e-invoices since January 2025, with issuing phased in from 2027. Belgium and Poland started their mandates in 2026, and France began on 1 September 2026. In each case, a PDF by email stops being enough for domestic B2B invoices.

Does it work with DATEV, Xero and QuickBooks?

Yes. Xero and QuickBooks Online have mature APIs and official MCP servers. DATEV integrations go through its marketplace interfaces, which require partner onboarding, and most German firms already exchange data with DATEV through standard export formats that the build can produce.

Also browse: Accounting firms · Xero · DATEV · QuickBooks · all use cases