Note: this is an illustrative example, based on typical financial automation projects. The data is representative and the company has been anonymized.
A distribution company, with several accounts at different banks and a high volume of daily receipts, automated bank reconciliation by combining robotic automation to extract transactions with ERP integration to record and reconcile the receipts. The process, which previously kept one person almost fully occupied during month-end close, now runs automatically, with human intervention only for exceptions.
The context
The company operated with several bank accounts and received hundreds of transfers a day, linked to customer invoices. The monthly close required manually downloading the statements from each online banking portal, identifying which invoice each transaction matched, and recording the reconciliation in the ERP. It was repetitive work, prone to error and concentrated on the highest-pressure days of the month.
The challenge
The process had exactly the conditions in which automation pays off:
- High volume: hundreds of transactions daily, with a growing trend.
- Clear rules: matching a transaction to an invoice follows defined criteria, such as amount, reference and date.
- Sources with no direct integration: the banking portals did not offer a simple way to export automatically, which made extraction manual.
- Impact on the close: the delay in reconciliation delayed the company's financial visibility.
The phased solution
Engibots implemented the solution in two phases, to reduce risk and show value early.
- Phase 1, automatic extraction: robotic automation bots access each banking portal, download the transactions for the period and organize them into a single format. This phase eliminated the manual collection of statements (see when RPA is the right choice).
- Phase 2, reconciliation in the ERP: an integration records the transactions in the ERP and automatically reconciles those that meet the defined rules, routing only the ambiguous cases to human review.
Splitting the work into phases made it possible to put the first part into production quickly and validate the gains before moving on to the integration (see systems integration).
Results
Illustrative results of this approach:
- Reconciliation time: reduced from several days a month to an automatic process, with the team handling only the exceptions.
- Errors: a significant decrease in matching errors, from eliminating manual transcription.
- Visibility: financial information available earlier in the close, with an impact on decision-making.
- Team: time freed up for analysis and control tasks, instead of data entry.
Lessons
Three useful conclusions for anyone considering a similar project:
- Start with extraction: automating data collection first delivers immediate gains and paves the way for integration.
- Design for exceptions: the value lies in automating the normal case and routing the exceptional case well, not in trying to automate everything.
- Measure before and after: quantifying the time and errors of the manual process is what makes it possible to demonstrate the return (see how to calculate the ROI of automation).