Accounting automation

Why accounting is a good candidate — and where the line sits

Most of the work in accounting is not posting entries but everything before them: fishing invoices out of mailboxes and portals, retyping data, working out what a document concerns and which account or project it belongs to, matching payments, chasing missing documents. That work is repetitive and document- and text-based — exactly the profile that automates best.

The line is equally clear: an AI model gives a probable result, and the books require a definite one. That is why, in the projects we run, the system never posts on its own — it prepares a proposal, and it is approved by the person who carries the professional responsibility. This is not caution for show: the human check reduces the risk of a silent error in the books. The pilot measures how accurate the proposals are and how long it takes to check them, so the value of the automation can be assessed.

What we actually automate

• Document intake. Collecting invoices from mailboxes, supplier portals and scans into one register, with duplicates recognised. The least glamorous stage and the most common source of chaos — which is why it usually comes first.
• Reading and posting proposals for approval. Reading the data off a document — including layouts seen for the first time — and proposing the posting from the history of similar documents: account, project, cost centre, VAT. The accountant approves, corrects or rejects; corrections are recorded and used in periodic adjustments to the system — that is what improves the proposals over time.
• Payment reconciliation. Matching bank statements to invoices, with a list of exceptions to explain — instead of searching two lists by hand. It automates well because confident matches are unambiguous, and anything uncertain goes to the exceptions list for approval — not straight into the books.
• Approval flow. Routing a document to the right person, reminders, and making sure nothing gets stuck — with a history of who approved what and when.
• Completeness at close. A list of the documents still missing for the month-end close, and the chasing correspondence drafted — instead of remembering in the last week.

Structured invoices from KSeF — Poland’s national e-invoicing system — change the balance of this work: the document arrives in a form the system reads directly, so reading from a scan disappears and the value concentrates in posting decisions, reconciliation and approvals. Less common documents outside KSeF, such as foreign invoices, receipts and contracts, still need reading.

What we do not promise

• Bookkeeping without a person. The system reads, proposes postings and reconciles, but the proposal is approved by the person who retains professional responsibility and signing authority. That is not a temporary limitation but the right division of roles: an AI model gives a probable answer, and the books require a definite one.
• Unattended operation from day one. Every stage — from posting proposals to the chasing correspondence — starts in a mode where a person approves every result. The decision to loosen oversight is made on pilot data, not by an assumption written into the offer.
• Automation that does not pay for itself. If the document volume is so small that the saving will not cover the system’s upkeep, we say so plainly in the first phase. The full set of conditions under which we advise against a project is collected on what we don’t do.

What happens to your accounting data

Accounting documents are the company’s financial data and its counterparties’ personal data, and payroll can additionally contain special-category data — health or trade-union details, for example. Before any document reaches a model, you have to settle which data leaves the company at all, where it goes and on what basis — from the choice of provider and processing region to limiting the data to the minimum a given step needs. How we approach it, and what to ask any supplier, is set out on our security and GDPR page.

How we work

As in every project: we pick one process and measure one indicator before we start, for example minutes from an invoice arriving to it being posted. Then we run the smallest possible pilot, with a person approving every result. Before switching over, the system runs alongside the current way of working for a while. The full course is set out in our implementation methodology, and the full scope of the service — from process audit through to upkeep — on our AI implementation page. The five conditions a process must meet to qualify for automation are set out on our AI automation page — accounting processes meet them more often than most, because they recur monthly and their data comes as documents.

What it costs

Cost is driven by the number of channels documents arrive through, the accounting system, and how reliable the posting proposals must be, that is, how rarely they may be wrong — not by the document count in itself. One intake channel and one system is a matter of weeks; five mailboxes, paper and two systems is a different project. All the components, upkeep included, are broken down in our note on what an AI implementation costs.

Frequently asked questions

Can AI post accounting documents on its own?

Prepare the posting — yes; take responsibility for it — no. The system reads the document, proposes the posting based on the history of similar documents, and hands it over for approval. Professional responsibility and sign-off stay with the accountant — and that is not a temporary limitation but the right division of roles: an AI model gives a probable result, and the books require a definite one. The value is that approving a well-prepared proposal takes seconds, while retyping and coding a document by hand takes minutes.

Will this work with our accounting system?

The audit at the start of the project settles that. We check whether the API (the interface other software uses to connect to the system) or file import supports the data and actions the process needs. We then confirm the integration’s feasibility, scope and cost, including whether an existing system feature is sufficient. Polish accounting systems vary widely here — from full APIs to file import only — and the cost of integration can decide whether the project is worthwhile. What often matters more than the system is how documents reach it: mailbox, paper, supplier portals — because that is the stretch we automate first.

What does KSeF change in accounting automation?

It shifts the balance of the project in automation’s favour. A structured invoice from KSeF — Poland’s national e-invoicing system — does not have to be read from a scan: the data arrives in a form the system reads directly, so the least reliable step disappears. The value moves to the later steps: posting decisions, assignment to projects and cost centres, reconciliation and approvals. Less common documents outside KSeF — such as foreign invoices, receipts and contracts — still need reading.

What does accounting automation cost?

Cost is driven by the number of channels documents arrive through, the accounting system, and how reliable the posting proposals must be, that is, how rarely they may be wrong. A company where invoices arrive through one channel into one system is a different project from one with five mailboxes, paper and two systems. Upkeep is a separate line: per-document fees and periodic retesting after model version changes. We break all the components down in our note on what an AI implementation costs.

Can an accounting office implement this too?

Yes — and offices are often a better candidate than a single company, because the same process repeats across dozens of clients, so the integration cost spreads over more documents. The difference concerns data: an office processes documents on its clients’ behalf, so the model provider’s part in that processing needs to be put in order in the office–client relationship as well. That is a question to settle before the project, not after it.