We are live · Workspace and Law pricing available
Optimaite

Documents & Structured Storage

4 min readUpdated May 27, 2026Auch auf Deutsch verfuegbar

Documents in Optimaite Business are organized through Structured Storage (Strukturierte Ablage), secured in the Vault (Tresor), enriched by AI field extraction, and validated through Reconciliation (Abgleich). Together these features give you full control over your document lifecycle.

Structured Storage (Strukturierte Ablage)

Structured Storage organizes documents into a hierarchy of groups, subgroups, and templates:

  • Groups — top-level categories (e.g. Invoices, Contracts, Purchase Orders)
  • Subgroups — subdivisions within a group (e.g. by year, supplier, project)
  • Templates — predefined document templates that set expected fields and filing rules

When AI processes an incoming document, it suggests a group and subgroup based on the document type and extracted content. Confirm the suggestion or move the document manually.

Upload or receive

Documents enter the system via inbox processing, direct upload, or API integration.

AI classifies and files

The AI pipeline determines document type, extracts fields, and suggests a filing location in the structured hierarchy.

Review and confirm

Verify the suggested location, field values, and entity assignments. Confirm to finalize filing.

Vault (Tresor)

The Vault is a secure storage area for sensitive documents that require restricted access:

  • Contracts with confidentiality clauses
  • Financial records subject to retention policies
  • Personal data subject to GDPR or industry regulations

Vault documents are encrypted at rest and access-controlled separately from standard structured storage. Administrators configure vault policies under Verwaltung → Tresor.

External sources

The vault can also ingest documents from external sources — connected cloud storage, SFTP servers, or partner portals. External sources are configured per group and synchronized on a schedule or on demand.

Document field extraction

AI extracts structured data from every document. Extracted fields fall into three categories:

Scalar fields

Single-value fields such as invoice number, date, total amount, due date, currency, and sender name. These appear in the document detail panel and are searchable.

Grouped fields

Related fields grouped together — for example, a vendor block (name, address, tax ID) or a bank details block (IBAN, BIC, bank name). Grouped fields keep related data organized.

Table / collection fields

Line items extracted as repeating rows — invoice line items, delivery note positions, or contract schedule entries. Each row has its own set of columns (description, quantity, unit price, total).

Field extraction accuracy depends on document quality and format. Always review extracted values before confirming, especially for amounts and dates.

Document templates

Templates define the expected structure for common document types:

TemplateTypical fields
InvoiceInvoice number, date, due date, vendor, line items, totals, tax
Purchase OrderPO number, order date, supplier, delivery address, line items
ContractContract number, parties, start date, end date, value
Delivery NoteDN number, shipping date, carrier, positions, quantities

Templates drive field extraction — when a document matches a template, the AI targets the template's expected fields for extraction. Administrators create and manage templates under Verwaltung → Vorlagen.

Reconciliation (Abgleich)

Reconciliation compares extracted document data against external data sources to identify discrepancies:

Select documents

Choose one or more filed documents to reconcile — for example, a vendor invoice against a purchase order.

Choose comparison source

Select the external data source: a previously filed PO, an ERP export, a bank statement, or a partner system feed.

Review discrepancies

The system highlights mismatches — quantity differences, price variances, missing items, or date conflicts. Each discrepancy is flagged with severity.

Resolve or escalate

Accept matches, correct errors, or escalate discrepancies for manual review. Resolved reconciliations are logged for audit.

Reconciliation results are only as reliable as the source data. Ensure external sources are current and complete before running comparison.

Next Steps

Was this helpful?