Automatically categorize your documents

Manual document sorting can represent between 30% and 50% of the total time spent on document processing within organizations.
Each employee must open files one by one, identify their nature (invoice, quote, contract, purchase order), and then send them to the right department or workflow.

A logistics operator in France using Koncile reported that manually opening each transport provider’s documents could take 2 to 3 hours per day per operator, with a significant error rate.

Koncile offers an automatic document categorization feature.
Whether you import a document manually or send it through the API, the tool identifies the document’s nature based on the templates you have configured in the platform.

Let’s take a simple example:

You receive invoices and purchase orders, but the files are not named correctly. You need to open each one to understand what it is, and only then extract the specific information related to that invoice or purchase order.

With automatic classification, each document is recognized before you even open it.
And with the extraction engine, you can choose exactly which fields you want to capture for each type of document.

Koncile categorization is based on a lightweight yet extremely effective pipeline.

• Transformation of image content into text

If the document contains no text (scan, photo, image), Koncile automatically applies an advanced OCR engine to retrieve all usable content. If it already contains text (native PDF, Word, email), OCR is skipped.

• LLM-based analysis

Once the text is available, Koncile uses a specialized LLM to understand the nature of the document and assign it to the correct category: invoice, purchase order, quote, contract, bank statement, certificate, etc.

• Incorporating user instructions

For more complex cases, Koncile relies on user-provided instructions to go further in the classification.

Examples: distinguishing invoices by supplier or customer, sorting delivery notes based on business criteria, or segmenting similar documents coming from different entities.

These custom rules act as hints for the model, which automatically adjusts the classification to match the real business context.‍

• Reliable and adjustable results

Thanks to this combination of OCR + LLM + business rules, Koncile correctly identifies almost all documents on the first pass while allowing fine-grained classification when needed.

Here are common use cases seen with Koncile customers:

Accounting & Finance‍

Accounting teams automatically sort documents received from suppliers (invoices, expense receipts, credit notes, bank statements). Specific data is extracted for each type: amounts, VAT, due dates, line items, unit prices. Mandatory fields are checked for compliance.

HR‍

HR teams automate sorting of onboarding documents (identity documents, degrees, bank info, certificates). Employee documents (contracts, amendments, payslips, attestations) are automatically categorized into the digital file. Expense reports are recognized by type (restaurant, hotel, transport) with automatic extraction.

Procurement & Supply Chain‍

Purchasing teams automatically sort the entire procurement cycle: purchase requests, supplier quotes, purchase orders, delivery notes, invoices. Logistics teams organize CMRs, bills of lading, customs documents, and certificates of origin. Quality documentation (certificates, attestations, audits) is collected and updated with no manual effort.

Healthcare‍

Healthcare organizations automatically sort patient documents: prescriptions, test results, reports, imaging, medical certificates. Reimbursement requests are processed using automatic recognition of care forms. Regulatory health documents (vaccination certificates, aptitude documents, mandatory checkups) are identified and sorted.

Real Estate & Rentals‍

Agencies and property managers automatically sort tenant applications: ID documents, income proofs, rent receipts, attestations. All rental documents (leases, inspection reports, diagnostics) are categorized by property. Sale-related documents (forms, diagnostics, notarial documents) are tracked and organized.