Skip to content
Butterfly
TürkçeEnglishDeutschFrançaisAzərbaycan日本語한국어中文
Request a Demo

Your documents are read and classified,
then archived after approval.

Intelligent document processing and enterprise document management.

DocumentFlow reads documents in your file sources, classifies them by type, extracts metadata by type and finds exact duplicates. Documents are validated against rules defined per type, with e-Devlet (Turkish e-Government) QR checks and external system checks where needed. Suspicious or sensitive documents go to a review queue; only documents that meet your rule are published to the target archive.

Let’s review your document flow together How does it work?
SharePoint and file sourcese-Devlet and rule-based validationDuplicate detectionHuman approval and decision trail
AllIn reviewReady to publishDuplicates
DocumentTypeLibraryStatus
Service agreementContractLegalReady to publish
Supplier correspondenceCorrespondenceProcurementIn review
Technical assessment reportReportProjectReady to publish
Service agreement (duplicate)ContractLegalDuplicate
Draft contractContractLegalUnsigned
1Source readDocument retrieved from the source folder
2Type identifiedContract, legal library
3Metadata extractedParties, document date, subject
4Duplicate checkContent hash compared
5Signature and versionSigned, final version
DECISIONPublish to archiveMeets the publishing rule
Confidence score
Certificates · Documents sorted by type, library and statusRepresentative DocumentFlow screen · sample data
Problem

Which document is the original,
and which is a copy?

Documents pile up over the years across different folders and shared drives. Multiple copies of the same contract, drafts and final versions sit side by side. Deciding which document should go into the archive means opening and reading them one by one.

Copies and versions

Copies of the same document sit in different places; it’s unclear which one is the original.

Manual classification

Determining a document’s type and metadata means opening and reading each one.

Uncontrolled archive

When drafts or unsigned versions slip into the archive, trust in the archive drops.

How does it work?

From source to archive
in four steps.

DocumentFlow delivers Butterfly's intelligent document processing (IDP) capability as a product: it reads the document, determines what it is, extracts the required information and decides by rule whether it goes to the archive.

  1. 1

    Connect to the source

    Documents in your SharePoint or file source are read; subsequent syncs process only changed documents.

  2. 2

    Classify

    Each document is sorted by type and library, and metadata fields defined for that type are filled in.

  3. 3

    Validate and review

    Each document is validated against its type's rule set: e-Devlet QR and barcode checks, external system checks, maximum document age. Duplicates are linked to the original; suspicious or sensitive documents go to the review queue, and the decision is written back to Butterfly.

  4. 4

    Publish to the archive

    Only documents that meet the publishing rule (e.g. signed and final) are sent to the target archive.

From the field

Subcontractor onboarding documents
are verified every 5 minutes.

At the Türkiye plant of an international cable manufacturer, a PoC was set up to automatically validate contractor and visitor entry documents, and tested in production with real documents.

Read the use case
  • 13 document types, each with its own rule setThe HSE team’s rules were defined as document types: verification method, maximum document age, and review requirement.
  • e-Devlet and external system verificationThe SGK employment notification was verified through e-Devlet by reading its QR code, with no human intervention.
  • Human approval for sensitive documentsTypes such as criminal records went to the DocumentFlow review queue by rule.
  • The decision was written back to ButterflyApproval or rejection was recorded along with the decision maker and notes. All four live end-to-end tests passed.
Publishing rule

Only documents that follow the rules
go to the archive.

The decision for each document is recorded step by step. Reviewers see which signal led to which outcome and the decision’s confidence score alongside the document; in ambiguous cases, a human has the final say.

Documents that go to the archive

  • Those that meet the publishing rule defined by the organization (e.g. signed and final version)
  • The single document marked as the original among exact duplicates

Documents that don’t go to the archive

  • Drafts and unsigned versions
  • Copies of the original document; these are linked to the original
  • Suspicious documents whose review is not complete
Capabilities

Everything you need
for document processing.

Incremental syncAfter the initial scan, only changed documents are reprocessed; a full scan runs when needed.
Type and library classificationDocuments are sorted according to the organization’s library structure; classification can be corrected from the review screen.
Metadata by typeEach document type has its own field set; fields are configured as date, list or text by type.
Duplicate detectionExact duplicates are found by content hash and linked to the original document.
Certificate verificationDocuments are validated automatically with e-Devlet QR and barcode checks, external system checks and type-based rule sets; sensitive types go to human approval.
Permission-based accessAccess to documents is granted according to Butterfly's permission model; documents work integrated with processes.
Which teams is it for?

Teams with many documents
and a critical archive.

Legal and contract teams

Legal counsel, contract managementWhy? ↻
Why?

If it’s unclear which version of a contract is in force, searching the archive offers no confidence.

What it does

Signed, final contracts enter the archive with the correct metadata, and drafts stay out.

Legal counsel, contract management

Quality and document control

Quality assurance, document controlWhy? ↻
Why?

In an audit, you need to be able to show why a document is in that library.

What it does

Controlled documents are classified by type, and every decision is traceable with its rationale.

Quality assurance, document control

Archive and records management

Corporate archive, information managementWhy? ↻
Why?

Migrating an archive without weeding out duplicates moves the problem into the new archive.

What it does

Documents in scattered folders are cleared of duplicates and moved into a single, organized archive.

Corporate archive, information management
Trust and control

People have
the final say.

DocumentFlow does not make decisions on its own in ambiguous cases. Decisions are recorded with their rationale, documents open according to permissions, and only documents that meet the rule go to the archive.

Control

  • Review queue for suspicious documents
  • Approval, rejection and reclassification
  • Step-by-step decision trail and confidence score
  • The publishing rule is set by the organization

Platform

  • Enterprise document management on Butterfly
  • Permission-based access control
  • Document flow connects to your own processes through App Builder and No-Code API Builder
  • Can be deployed on-premises
Needs assessment

How well does DocumentFlow fit your document flow?

8 questions, two minutes. At the end, see DocumentFlow’s fit scorecard for your document flow and the reasons behind it.

  • Document volume
  • Classification
  • Archive control
  • Audit
8 questions · about 2 minutes
Document volume1 / 8
Where are your documents stored?
Document volume2 / 8
Do copies or different versions of the same document sit side by side?
Classification3 / 8
How are document types and metadata determined today?
Classification4 / 8
Do you need to keep separate metadata fields per document type?
Archive control5 / 8
Is it a problem when drafts or unsigned versions get into the archive?
Archive control6 / 8
Do you have a written rule for which documents enter the archive?
Audit7 / 8
Do you need to show why a document is in a particular library or archive?
Audit8 / 8
Should access to documents be restricted by person or role?
See your DocumentFlow readiness scorecard · 2 min
Frequently asked

Quick answers
about DocumentFlow.

Which sources does DocumentFlow read documents from?
It reads documents from SharePoint and file sources. After the initial scan, only changed documents are reprocessed.
How are document types determined?
Each document is classified by type and library. Ambiguous cases go to the review queue, where a user confirms or changes the classification.
How are duplicates found?
The document’s content hash is compared. Identical documents are grouped and a single original is marked for the group.
Which documents go to the archive?
Only documents that meet the publishing rule defined by the organization, such as signed and final versions. Drafts, unsigned versions and duplicates are not sent to the archive.
Can we see the reason behind a decision?
Yes. The decision for each document is recorded step by step; which signal led to which outcome and the decision’s confidence score are visible alongside the document.
Can document validity be verified?
Yes. A validation method is defined for each document type: e-Devlet barcoded documents are checked against the official service by reading the QR code, and documents with a report address are checked against the external system. Types that require human approval by rule go to the review queue.
Can it be deployed on-premises?
Yes. DocumentFlow runs on the Butterfly platform and can be deployed on-premises; access is granted according to Butterfly's permission model.

Let’s review your document flow through one library.

Let’s look together at the source where your documents live, the document types that should enter the archive and your publishing rule. We’ll do the first assessment on a single library.

Let’s assess your document flow Explore the Platform