Skip to main content
Request Demo

Document AI · Classification

Automatically classify, split, and route every document.

Artificio identifies what each incoming document is, determines which Data Group it belongs to, separates multiple documents inside the same PDF, and routes each document to the correct extraction model, with no traditional training project required.

Request a demoExplore SAP Automation
  • No labeled classification dataset
  • Standalone or before extraction
  • SOC 2 Type II
  • ISO 27001
INBOUND · MULTI-DOCUMENT.PDFCLASSIFYING
  1. INInvoice · Pages 1–3Data Group: Supplier Invoice → Extraction ModelROUTED
  2. POPurchase Order · Pages 4–6Data Group: Customer PO → Extraction ModelROUTED
  3. ?Unclassified · Pages 7–8Data Group: Others → Human ReviewREVIEW
  • Classify invoices, POs, inspection reports, contracts, forms, and custom document types
  • Detect where one document ends and the next begins inside a combined PDF
  • Route each classified document to the right active Data Group and extraction model
  • Unmatched or uncertain documents become Others and stop for human review
No trainingNo labeled classification dataset
Page-awareDetect first & last page
Auto-routeTo the correct extraction model
Human-safeUnknown documents stop in Others

How it works

Classification before extraction, or used on its own.

When documents arrive from email, uploads, folders, APIs, or other channels, Artificio first determines what each document is. It can stop after classification or continue automatically into the correct data extraction workflow.

  1. 1

    Receive

    Email, upload, folder, API

  2. 2

    Classify

    Identify type and boundaries

  3. 3

    Assign Data Group

    Invoice, PO, COA, Others...

  4. 4

    Route

    Send to active model

  5. 5

    Extract

    Create structured record

Artificio Data Groups

The classification label is also the destination.

In Artificio, a Data Group represents the business document structure the system understands. Each Data Group contains the fields and tables associated with that document type. Classification decides which Data Group an incoming document belongs to and routes it to the active extraction model for that group.

Example · Supplier Invoice

Data Group: Supplier Invoice

An incoming invoice is assigned to the Supplier Invoice Data Group before extraction.

invoiceNumber
text
invoiceDate
date
vendorName
text
purchaseOrder
text
lineItems
table
Example · Customer PO

Data Group: Customer Purchase Order

The same inbound channel can contain a purchase order, which routes to a completely different extraction structure.

customerName
text
customerPONumber
text
requestedDate
date
shipTo
address
orderLines
table

No classification training project

Classification works with the document intelligence you already configure.

The classification step is designed to work from the active document/Data Group definitions available in Artificio instead of requiring a separate labeled classification dataset and training cycle.

  • No labels

    No classification dataset to maintain

    Avoid a separate manual labeling project just to teach the router what an invoice, PO, or inspection report looks like.
  • Dynamic

    Routes against active Data Groups

    The available classification destinations align with the business document structures configured in the platform.
  • Controlled

    Unknown stays unknown

    If Artificio cannot determine a reliable match, it sends the document to Others rather than forcing an incorrect route.

Common scenarios

Useful anywhere documents arrive mixed together.

Use classification alone for intelligent sorting and routing, or as the first step before data extraction.

  • Shared mailbox

    Email document routing

    Classify invoices, POs, statements, forms, and other attachments and send each to the right processing path.
  • Batch scanning

    Mixed scanned packets

    Separate mixed batches without requiring staff to manually split every file before processing.
  • Multi-document PDF

    Repeated documents

    Several invoices or POs inside one PDF become separate logical documents and independent records.
  • Manufacturing

    Quality document packages

    Separate COAs, inspection reports, specifications, and supporting documents before the quality workflow begins.
  • Finance

    AP intake

    Distinguish invoices, credit memos, statements, and supporting documents before extraction and matching.
  • Standalone

    Classify without extraction

    Use the capability simply to identify, separate, tag, or route documents.

Automating something that isn't on this page? The agents adapt to new document types, new systems and new rules. Tell us what you are automating and we will show you the nearest thing we run.

Talk to our team

Document boundary detection

One PDF can contain many documents. Artificio separates them automatically.

A sender or scanner may combine several logical documents into one file. Artificio can identify where each document starts and ends, classify each segment independently, and create a separate record for each document.

Five invoices in one PDF can become five invoice records. A mixed packet containing purchase orders, invoices, inspection reports, and other pages can be separated and routed to different Data Groups.

  • Invoice A · Pages 1–2INVOICE
  • Invoice B · Pages 3–4INVOICE
  • Purchase Order · Pages 5–7PO
  • Inspection Report · Pages 8–10QUALITY

Document types

Classify the documents your business already receives.

InvoicesPurchase OrdersCredit MemosStatementsContractsFormsCertificates of AnalysisInspection ReportsPacking ListsBills of LadingRemittancesClaimsApplicationsCustom Document Types

What happens next

Classification is the routing layer before extraction.

Once Artificio knows what the document is, other Platform capabilities can take over.

FAQ

Common questions about AI document classification.

Does Artificio require training data to classify documents?
No traditional labeled classification dataset is required for the workflow described here. Artificio classifies against the active document/Data Group definitions configured in the platform.
What is a Data Group in Artificio?
A Data Group represents a business document structure and the fields and tables associated with that document type. Examples include Supplier Invoice, Customer Purchase Order, and Certificate of Analysis.
Can Artificio classify several documents inside one PDF?
Yes. Artificio can identify logical document boundaries, separate the segments, classify them independently, and create separate records.
What happens if a document cannot be classified?
It is assigned to Others and processing stops for human review rather than forcing the document into an incorrect extraction path.
Can classification be used without extraction?
Yes. It can be used as a standalone capability for identifying, splitting, tagging, and routing documents.

Security & compliance

Enterprise security across every solution

ISO 27001:2013 certified, SOC 2 Type 2 compliant, GDPR and HIPAA ready. Every agent action is logged, auditable, and runs in isolated environments.

  • ISO 27001:2013
  • SOC 2 Type II
  • GDPR ready
  • HIPAA ready

See it on a mixed document package

Send one PDF. See Artificio separate and classify it.

Bring a PDF containing invoices, purchase orders, inspection reports, forms, or other business documents. We'll show how Artificio identifies document boundaries, assigns Data Groups, and routes each document to the right next step.

Request a demoTalk to sales

SOC 2 Type II · ISO 27001