Why automating packing lists looks attractive
Packing lists are repetitive. They always come from the same suppliers, always contain the same columns, and processing them is pure manual work: copying lines into your WMS, ERP, or inspection software. For a company handling hundreds or thousands of packing lists a year, that is a clear candidate for automation. The volume is high enough, the process is defined, and the added value of manual data entry is zero. HDG, active in food and quality inspection, processed 200,000 packing lists per year and eliminated 90% of the manual work after automation was introduced. That sounds like a smooth transition, but the picture is only partly accurate.
What goes wrong when you only build for the standard case?
Most packing lists arrive neatly as PDFs or Excel files with a fixed structure. The automation process is built around that. But then a supplier sends a packing list as a screenshot from their own system, or another party uses a slightly different column order, or a document arrives as an attachment in the wrong format. Anyone who has not designed an exception process has a problem: the document is not processed, no one receives a notification, and only later does it become clear that something slipped through the cracks. The lesson is simple and yet keeps being learned the hard way: automation works for the standard case. But your process stands or falls on how you handle the exceptions.
How do you design the exception process correctly?
Three principles help. First: make deviations visible, not invisible. A document that could not be processed automatically must immediately land in a queue that someone actually monitors, not disappear into a folder nobody opens. Second: have a person review uncertain cases. Not as a bureaucratic step, but as a genuine safety valve. When a system cannot read a field with sufficient confidence, it routes the document back for human review. That is not a failure of automation; that is the system working correctly. Third: track which exceptions recur most often. If one supplier consistently sends documents in a non-standard format, that is input for a targeted adjustment, not a reason to question the entire system.
Is the quality of your source document also a factor?
Absolutely, and this is the point where honesty pays off. When packing lists are structurally poor, think handwritten additions, partially printed PDFs, or documents scanned at an angle, automation delivers less value until that problem is resolved. No system reliably extracts data from a document that is unreadable to a person with good eyesight. The question you need to ask upfront is not only: how many packing lists come in? But also: how many of them are consistent enough in quality to be read reliably? The answer to that second question largely determines what you can expect from automation.
What does automating packing lists actually deliver when it is set up correctly?
When the volume is high enough, the process is defined, and the exception handling is in place, most of the manual entry disappears. Employees who previously spent their time copying lines into inspection software or a WMS are freed up for the work they were hired to do: checking, flagging, deciding. At HDG, that meant a freed-up FTE returning to the inspection work itself. That is the real gain, not the time saved on a single document, but the structural shift in what people do with their workday. The lesson that keeps coming back: start with high volume and good document quality. Expand from there. Anyone who tries to automate the entire document landscape at once will get stuck on exceptions they did not anticipate.