The Files That Arrive and Get Processed by Hand
Last updated:
The pattern is everywhere
A supplier emails a file every morning. Someone opens it, checks it, reformats it and loads it somewhere. Or a folder receives files that someone processes when they notice.
It is repetitive, rule-based and error-prone, which makes it close to ideal for automation.
What the automation does
- Watches the mailbox or folder for new files
- Identifies what kind of file it is
- Validates the structure and content before processing
- Processes it into the destination
- Archives the original and records what happened
The validation step is what makes it safe. A file with a changed column order processed blindly corrupts data in a way that is difficult to unpick.
Handle the awkward realities
- Files that arrive late, or twice, or not at all
- Column order or naming changing without notice
- Encoding differences that break accented characters
- Zero-byte or partially written files
- The occasional file that is a different format entirely
Alert on absence
A file that does not arrive is as important as one that fails. If the daily file has not appeared by the expected time, somebody should be told.
That is the alert that catches a supplier changing their process without telling you, which happens more often than anyone expects.
Keep the originals
| Keep | Why |
|---|---|
| The original file | Reprocessing and disputes |
| A processing record | What happened and when |
| Anything quarantined | Diagnosis and reprocessing |
| The validation result | Evidence the check ran |
| A retention period | So it does not grow indefinitely |
Frequently asked questions
Can it read email attachments?
What if a file is malformed?
How do we handle files that arrive twice?
What does this cost to build?
A file that arrives every morning and gets processed by hand?
That is a contained automation with a clear daily saving. Happy to scope it.
Related services
What we build for problems like this one