On the first working day, your Aivell opens the close checklist, pulls the month's figures from your database, runs the reconciliation checks, mails each task owner what they owe and chases until it is confirmed done. Small differences are noted, big ones escalated, and the closing summary waits for the controller's sign-off.
starts when: The first working day of the month, on schedule
Every close is the same fifteen steps, and every close they are run from memory. Pull the figures. Reconcile the bank. Wait for the fuel card statement. Ask sales about the credit note. The steps live in a spreadsheet nobody updates and in the head of the one person who has done it forty times — and the close stretches to day ten because step six waited three days for a reply nobody chased.
Most of those steps are extraction, comparison and reminding: mechanical work wrapped around a handful of true judgment calls. Analysts put around 45% of professional-services work within reach of current automation — the close is where that estimate feels conservative.
Underneath, the guardrails are unambiguous: the books are read-only to the AI, the workbook is the only file it writes, the recipient list cannot grow. The checklist runs the same way every month — which is, after all, the entire idea of a close.
No flowcharts, no code — a chore is just a message to your Aivell, in your own words. It shapes it into a solid, guarded procedure and follows it to the letter. This one:
on the first working day, open a fresh close checklist from the template
pull the month's figures from the database into the close workbook
run the reconciliation queries — bank against ledger, sales against invoices
differences under €50 go on the notes tab; anything bigger goes straight to Franca
email each task owner their open items, and again every morning until confirmed
tick a task only on the owner's confirmation, never on silence
on day 5, send the status summary — done, stuck, and who we're waiting on
the closing email to the accountant waits for Franca's sign-off
Connected with single-click passkeys — no copied tokens, no OAuth hell.
Every tool can be limited to exactly what this chore needs. For this one, you might set:
It can't — the database guardrail is read-only, which is a property of the connection, not a promise. It finds the €312 gap between bank and ledger and brings it to Franca; a human decides what gets booked, and books it.
That is what the checklist is. "Ask Davide to confirm the fuel card statement" is a perfectly good task — the Aivell mails Davide, chases him each morning, and ticks the item when he confirms. If a human can be briefed on it, it fits.
On the box, in your office. The queries run against your own database over your own network, the workbook lives on your own storage, and the reasoning happens onboard. Your P&L never rides through anyone's cloud.
A reminder app tells people to do things. Your Aivell does the pullable work itself — extracts, reconciliations, the status report — and only chases people for the parts that are genuinely theirs. It also refuses to tick a box on silence, which is where checklist apps quietly lie.
Your Aivell compares spreadsheets with the database, explains every mismatch in one line, and mails the list — financial data checked behind your firewall.
Your Aivell queries the database, builds the weekly report and mails it out on schedule — computed onboard, so your numbers stay behind your firewall.
Hours chased, checked and compiled into one master sheet before the payroll cutoff — by an AI colleague that works entirely inside your office network.
An on-premise AI colleague for accounting firms: reads invoices, chases client documents, minds the tax calendar. Client files never leave your office.
An Aivell is an on-premise AI colleague: a small box that plugs into your network, runs its own AI onboard and takes the repetitive work off your desk. Prompts, documents, data — nothing ever leaves your office. Unplug the internet: it keeps working.
You describe each task as a chore, in plain language. Your Aivell turns it into a solid, guarded procedure and runs it in the background — with guardrails on every tool, a complete audit trail, and approvals in your hand when you want them. One-time setup, fixed monthly fee. No tokens, no overages.
onboard inference · air-gapped · fixed monthly fee