Supplier invoice intake
Finds supplier invoices in a mailbox, uses AI only for uncertain documents, and uploads valid ones to the accounting system.
Overview
Invoices arrive in a mailbox among many other emails and documents. This automation first sorts them using rules and uses AI only for documents that the rules cannot classify. It extracts invoice numbers, uploads valid invoices in batches to the accounting system, and logs everything for review.
What has changed
- Reduced manual mailbox management
- Invoices are consistently selected and validated
- Logs and notifications show what has happened
At a glance
In één oogopslag
Overzicht
Invoices arrive in a mailbox among many other emails and documents. This automation first sorts them using rules and uses AI only for documents that the rules cannot classify. It extracts invoice numbers, uploads valid invoices in batches to the accounting system, and logs everything for review.
Wat er verandert
- Reduced manual mailbox management
- Invoices are consistently selected and validated
- Logs and notifications show what has happened
How it works.
What it does per email
- Downloads the previous day's emails and filters them using rules
- Uses AI to classify the remaining uncertain documents as invoices or not
- Extracts invoice numbers and sender details
- Uploads invoices in batches of 20 and verifies the counts
- Moves processed emails to the appropriate folders
- Emails the run log and alerts a developer if any errors remain
Rules
Rules take precedence: known non-invoice senders, keywords in subjects, filenames and text, and a final check of filenames. Only documents that the rules cannot classify go to AI, which determines whether they are invoices. Everything thereafter, including invoice numbers and uploads, follows fixed logic.
Check out other automations.
Frequently asked questions.
No. The digital colleague works in the screens and systems your team already uses, whether they are legacy applications without an API or modern ones, so there is nothing to replace.
Yarado builds it, runs it and monitors it. When a screen or a system changes, we adjust the automation, so your team needs no developer or automation skills of its own.
Yes. Every run is recorded in the Yarado portal, where you can see which processes ran, when they last ran, what they produced and what went wrong. Depending on the automation, the processing log is also emailed to the team that needs it.
Yarado monitors every run and is alerted automatically when one fails, so a problem is picked up on our side. Cases that need a person are reported to your team with the reason, in the portal or by email.
Yes. Automations are built around your own steps and systems, and this page is just one example of what is possible. A first automation is typically live in weeks, not months.