Lieferschein-Automatisierung
Intrapass receives dozens of delivery notes from their partners every month, all of which need to be matched against orders and turned into invoices. We automated the process with an AI that knows the specific quirks of their operation.
At Zurich Airport, Intrapass offers a range of services, including Baggage Delivery at Lost and Found. After every delivery, the signed delivery notes from their partners come in, on average 40 to 60 per month. These had to be matched manually against orders, checked, and turned into invoices. Automation was an obvious idea but never quite worked, because delivery notes are not standardised. Every supplier has their own format, plus exceptions and quirks that only someone who has run the operation for years truly knows. That kind of task was never really within reach of classical automation. With AI, it now is.
We built an AI workflow that knows Intrapass's quirks and runs the necessary checks cleanly. Incoming delivery notes are read, matched against the corresponding orders, amounts and line items verified. What the AI can handle with high confidence, it handles. What stays unclear goes to the responsible team member with a concrete clarification suggestion, fully contextualised, instead of a generic error message. For every invoice, the AI generates a summary, a short explanation of what was processed and how. That keeps the process traceable, verifiable at any time, and gives staff confidence in what the AI is doing. The finished invoices land as drafts directly in Bexio, where they get one more pass before being released. Important: we did not have to change Intrapass's folder structure or introduce new tools. The workflow sits on top of their existing processes, not next to them.
Monthly effort has shrunk to about 15 minutes. Delivery notes are fully checked, discrepancies are flagged with clear context, invoices are ready by the time someone opens them. For an SME, this kind of small, targeted integration is gold today. It can be implemented quickly, unlocks meaningful time savings even when it seems modest at first glance, and can be tailored precisely to the operation's own workflows and exceptions. Instead of adapting to a standard tool, the business gets a tool that understands it. Made possible by AI as a development accelerator.
“We now have a workflow that fits our operation, not the other way around. What used to take a lot of time is done in fifteen minutes today.”