What is documentclassification?
Document classification assigns each incoming document to a type you define (invoice, contract, claim form), so the right extraction and routing can follow.
The sorting step everything depends on.
Classification answers the first question of every document workflow: what is this? The answer decides which schema extracts it, which checks apply and where results go. Done manually, it is the person who opens every file; done in software, it is a classifier assigning each document to one of your types.
Modern classification layers several methods: fast rule-based checks on filenames, content-based checks against expected structures, and semantic evaluation by a language model for the cases rules cannot separate. Ordered as a chain, cheap checks run first and expensive ones only when needed.
Your types
Classes are defined by you, not by a vendor catalog: as niche as your documents are.
Chained methods
Filename rules, structure checks and semantic LLM evaluation run in order until one matches.
Confidence handled
Documents no classifier places confidently go to review instead of being guessed.
Classification in MiruIQ.
MiruIQ pipelines classify with a configurable chain (file-name, single-structure, multi-structure and semantic classifiers) against your own types; an exclusive mode can even pick the right schema for mixed inboxes automatically.
Document classification questions
What teams ask about classification.
No. In MiruIQ, classes are whatever you define: standard types like invoices work out of the box conceptually, but a document class specific to your firm classifies just as well, because the semantic classifiers evaluate your definition rather than a fixed catalog.
The document parks in the human review queue instead of being forced into the best-scoring wrong class. A reviewer assigns the type in one click, and recurring uncertainty tells you precisely which class definition or rule to sharpen.
See it on your own documents.
The 14-day free trial covers 200 documents: enough to define your schemas, run real files and measure the results before anyone calls you.