Document classification with AI, without a template per sender.
Every mailroom holds dozens of document types across hundreds of layouts. Parashift recognises the type per page by content, with over 400 preconfigured document types and your own types without programming.
Sorting is manual work before anything is read.
Before a document can be processed, someone has to know what it is. In most organisations a person does that: leafing through stacks, inserting separator sheets, assigning documents to the right queues.
Template-based systems only help while sender and layout are known. Every new invoice layout, every new form means a ticket, a rule, an approval. The maintenance grows faster than the benefit.
What the classification does.
400+ preconfigured document types
Invoices, delivery notes, contracts, forms, identity documents and official post are ready to use. Activate instead of train.
Your own types without code
Describe the type, check it against examples, activate it. No development ticket and no rule set per sender.
Classification per page
A stack is not filed as a whole. Every page gets its type, so mixed inbound breaks apart cleanly.
Confidence per decision
Every assignment carries a confidence value. You set the threshold for automatic processing.
Human in the loop
Below the threshold the case goes to the review interface instead of into the booking. Built for mailroom and BPO teams.
Document Swarm Learning®
The model learns from corrections across the whole platform. New layouts are often known before they reach you.
From stack to sorted inbound.
Any format
PDF, scan, fax, email attachment or portal upload, in any order.
Document boundaries
The stack is split into individual documents by content, without separator sheets or barcodes.
Type per page
Every page receives its document type together with a confidence value.
Right department
The item goes to the queue, workflow or target system, unclear cases go to review.
Template-based classification and Parashift.
| Template-based | Parashift | |
|---|---|---|
| New layout from a known sender | Adjust template, test, release | Read without intervention |
| New sender | Create a new template | No setup needed |
| Mixed stack | Separator sheets or barcodes required | Boundaries detected by content |
| Handwriting and fax quality | Frequent failure | Part of normal operation |
| Maintenance effort | Grows with every sender | Configuration per document type, not per sender |
| Time to first result | A project over weeks | First run on the same day |
What changes in the mailroom.
preconfigured document types out of the box
Parashift platformdocument types running at large accounts
Parashift customer baseless effort in input management
Parashift project experienceClassification is one building block.
Data extraction, document separation and the digital mailroom build on it: first know what a document is, then read it and hand it on.
Document classification, answered briefly.
What does AI classification do differently from classic methods?
How many document types are available?
How do I create my own document type?
What happens with an uncertain assignment?
Does this work on mixed stacks without separator sheets?
How does the result reach our systems?
Classify documents automatically.
Book a no-obligation briefing: we will show the classification on your own documents.
Reply the same working day · no sales pressure
