Bank Reconciliation
Overview
Takes the payments and receipts the ERP could not reconcile and confirms each one against the real bank statement. It pulls the ERP's flagged paid-bill exceptions and open invoices awaiting cash, then signs in to the bank, searches the operating account by reference and reads what actually settled - quantifying the four ways money and ledger drift apart: a missing payment, a transposed amount, a double payment and unapplied cash. Writes a reconciliation CSV. The story is evidence: the ledger raises the question, the bank statement answers it, and every discrepancy comes out with a number attached.
Bank Reconciliation
The ledger says a bill was paid. The bank is where you find out whether it actually was — at the right amount, once, and only when it should have been. Reconciliation is the control that catches the gap between the two, and it is exactly the kind of patient, evidence-gathering work a robot does better than a person at month-end: for every item in question, pull the statement, read what settled, and put a number on the discrepancy.
This flow does that. It takes the payments and receipts the ERP could not tie out and confirms each against the operating account, classifying the four ways money and ledger drift apart.
It spans two fictional training systems — RAP One, the ERP, and AcmeBank, the bank — whose data is entirely synthetic and reconciles to the cent.
What it produces
bank-reconciliation.csv in your home folder: every flagged item with its verdict and the exact
figures from both sides. On the seeded data that is five confirmed discrepancies:
- Missing payment —
NW-2026-0442is Paid in the ledger but no debit left the bank; €4,998 outstanding. - Transposition —
HE-2026-06-4471settled at €1,824.33 against a ledger amount of €1,842.33 (a transposed digit, €18.00 out). - Double payment —
VF-2026-06-8801was paid twice: two debits of €489.90. - Unapplied cash —
INV-2026-0641(€8,901.20) andINV-2026-0655(€6,330.80) were received in the bank but the invoices are still open in the ledger.
How it works
1. Pull the ledger's open questions
Sign in to the ERP. Its exceptions worklist already flags the paid bills it could not tie out, and the customer invoices where cash is expected but the item is still open. The robot reads the reference and amount for each — the items to prove against the bank.
2. Prove each against the bank
Two-factor sign in to the bank and open the operating account. For every flagged item, search the statement by its reference and read what actually settled — no debit, a debit at the wrong amount, two debits, or a credit that arrived. The bank statement is the independent record; the count and amount decide the verdict.
3. Classify & report
Zero debits is a missing payment; one debit at the wrong amount is a transposition; two debits is a double payment; a credit against a still-open invoice is unapplied cash. Each verdict carries the figures from both sides, written to a CSV a finance team can action.
Running it
Ready to run as-is. It signs in with the published training credentials (HTANAKA for the ERP,
hiroshi.tanaka@globex.example with a static one-time code for the bank), which are not secret. Real
system credentials belong in the Robomotion Vault, never in a flow.
More web automation templates
See all 18 →- Payment Run with Hold ReportRuns the weekly vendor payment run in a SAP-style ERP and then accounts for every bill it refused to pay. Signs in with credentials read from the Robomotion Vault, proposes the bills due within seven days, confirms the run so a journal entry is posted per document, and verifies the posted summary against the proposal before going on. Then it opens the bills list, filters to the blocked bills, and reports each one with the reason it is on hold - a pending goods receipt, a quantity or price mismatch against the purchase order, or a missing PO reference. Writes a two-section CSV covering what was paid and what was held back, with the value held back totalled.
- Shipment Status ReportSigns in to a carrier portal and sweeps every page of the shipments list into one report, then narrows that sweep to the shipments that need a person. Reads the login from the Robomotion Vault, walks the pager with a Label and a GoTo until the Next control is disabled, and writes two CSVs: the full book, and an exceptions file covering customs holds, failed deliveries and delays, oldest booking first, with the billed value stuck in each one. Closes with a Log line carrying the counts and the total.
- Logistics Exception CenterBuilds one worklist of every active delivery exception from the carrier, then dedupes it against the systems that may already be handling it - the mailbox that was notified, and the help desk where a ticket may already exist. Joining the three on the tracking number tells the robot which exceptions were mailed, which already have a ticket, and which nobody has picked up yet. Writes a consolidated exception CSV. The story is consolidation: three systems each hold part of the picture, and the robot joins them so a stuck shipment is neither missed nor double-worked.
- Monday Morning BriefingOne robot, every system, the whole week's work on a single page. It signs in to the CRM, the mailbox, the ERP, the carrier and the help desk in turn, reads the one number that matters from each - duplicate contacts, invoices to post, ledger exceptions, stuck shipments and refund requests - and writes a consolidated briefing with a total across all five. The story is reach: the back-office queues a team would open one browser tab at a time, gathered into one report before anyone has had coffee.
- CRM Duplicate CleanupCleans up a CRM by merging duplicate contacts - the same person entered twice under name variants like Bob and Robert. Signs in, switches the contact list to its duplicates view, and merges the first remaining pair until none are left, keeping the primary record and folding each variant into it. Writes a before-and-after report. The "your CRM cleaned itself while you slept" demo.
- E-Invoice Portal HarvestDownloads every e-invoice for a tax period from a government portal and builds an index spreadsheet of what was collected. Signs in with a one-time code, walks the paged register until the portal says there are no more pages, then opens each document to read its government reference number and pull down both the XML and the PDF. Every download is gated behind a confirmation dialog and an "I am not a robot" checkbox.