How Stralevo Indexes 50+ Document Formats Into a Single Queryable Intelligence Layer
Forty-three percent of mid-market finance teams spend more than two hours per week locating documents that exist in multiple systems in incompatible formats — not analyzing the data, just finding it. The invoices are in the AP system. The contracts are in legal document management. The purchase orders are in the ERP. The delivery confirmations are in an email archive. The intelligence exists. The format fragmentation makes it unreachable.
When the CFO asks "what did we spend with Supplier X across Q3?" the answer currently requires sending that question to three different teams, waiting two days, and receiving back numbers that don't reconcile with each other. Each team pulled from the system they own. Each system knows a different slice of the same financial reality. Stralevo indexes all 53 format variants into one queryable layer — and makes that question answerable in 14 seconds.
Format indifference, not format standardization
Twenty years of ERP and accounting software have trained finance teams to believe the goal is standardization: get everyone sending invoices in the same XML schema, and the fragmentation problem disappears. That approach has been failing for thirty years. EDIFACT was going to become universal in the 1990s. XML was supposed to replace it in the 2000s. FacturX — France's national e-invoicing standard — is supposed to resolve it in the 2020s. Each standard reduces some format diversity, and new formats emerge. Format standardization requires supplier cooperation, contract renegotiation, and years of transition. Format indifference requires one intelligence layer that accepts everything already arriving in your system.
Six formats. Your accounting software reads two. The other four are invisible to your intelligence — until now.
European finance teams today receive documents across eight format families: structured documents (XML schemas including FacturX and UBL — the EU's Unified Business Language standard for electronic documents), semi-structured (PDF and Word), tabular (CSV bank statements and Excel attachments), image formats (TIFF scans, PNG, JPEG), EDI standards (EDIFACT, used extensively in logistics and retail), email-embedded content, HTML invoices, and proprietary ERP exports. Standard accounting software handles structured XML and high-quality PDF from large suppliers — the 30–40% of invoice volume that arrives in clean, machine-readable form. Stralevo was built for the remaining 60–70%: scanned PDFs from small regional suppliers, Excel invoices from service providers, EDIFACT messages from logistics partners, payment confirmations embedded in plain email text.
Inside a format-agnostic ingestion pipeline
SightCapture ICR — Stralevo's document ingestion engine — handles 53 distinct format variants through dedicated extraction engines for each format type. For image formats like TIFF scans, it applies optical character recognition and compares extracted results against expected field patterns, flagging fields where confidence drops below threshold rather than silently passing through incorrect data. A batch of 500 scanned invoices with handwritten corrections achieves 87% automated extraction at full confidence; the remaining 13% are flagged for manual review. Competitors without confidence scoring pass the full 100% through — including the portion containing errors.
For structured XML and JSON, a parser extracts explicitly labeled fields with near-100% coverage. For tabular formats like CSV and Excel, a dedicated extractor reads columns and maps them to a unified financial schema. When a logistics team sends purchase orders as Excel attachments, suppliers send invoices as scanned PDFs, and the ERP exports transactions as XML, Stralevo reads all three simultaneously and merges every extracted field into one searchable index. One query reaches all three at once.
One claim Stralevo doesn't make: 100% extraction accuracy from every document in every format, every time. Handwritten corrections on scanned invoices, ambiguous date formats across country conventions, and vendor-specific layouts that deviate from standards all create edge cases. The confidence scoring exists precisely because the honest answer to "how accurate?" is "field-by-field, and we'll tell you which fields to verify."
FacturX creates a compliance mandate — and a new format challenge
FacturX — France's national e-invoicing standard, mandatory for French companies above 250 employees from January 2026 — is often described as the solution to format fragmentation. It is also, technically, a new format challenge. FacturX is a hybrid: a standard archival PDF with a structured XML data file embedded inside it. Accounting software that reads only the PDF layer misses the structured data embedded within. Software that reads only the XML misses the human-readable PDF content and metadata. Full FacturX intelligence requires reading both components simultaneously and merging their fields into a single record. Stralevo handles FacturX's dual-component structure natively — which matters for companies receiving both incoming FacturX from large suppliers and legacy PDFs from the long tail of smaller vendors, a mix that will persist for years after the 2026 deadline.
Building format-agnostic intelligence now means being ready for every stage of the transition rather than rebuilding each time a new standard becomes mandatory.
Finding what format gaps hide
French tax audits reveal exactly what format gaps cost. When the DGFiP (France's tax authority) issues an audit notification, the company has 15 days to produce a FEC — the Fichier des Écritures Comptables, the standardized accounting export that must capture every transaction from every source. If EDI files from logistics partners were never indexed into the main system, or if scanned invoices were processed manually and not captured digitally, the FEC is incomplete. Minimum penalty for non-compliance: €5,000 — and an incomplete FEC typically expands the audit scope rather than closing it.
Renault processes more than 2 million supplier invoices per year across EDI, PDF, and portal uploads — historically routed to different teams because no single system could index all three simultaneously. For any mid-market company running Sage or Cegid alongside a logistics platform, the same split applies: financial formats in the ERP, operational formats in the supply chain system, each team working from a different slice of the same data.
A Stralevo client queried all invoices from the Bouygues supplier group across subsidiaries — spanning PDF invoices from 12 subsidiaries, XML purchase orders from the procurement system, and CSV payment confirmations from the bank. Results came back in 14 seconds, with every source cited. The same request in a format-siloed environment required contacting four teams and waiting two days for a result nobody could fully verify for completeness.
Price escalation hides specifically in format diversity. A supplier raising prices on specific line items by 2% per quarter can avoid detection for years when PDF invoices and EDI transactions are analyzed separately — the pattern only surfaces when all formats are queried together. Finance teams that detect this pattern find between €50,000 and €200,000 in recoverable overcharges per year in supply chains above €5 million in annual spend.
2026's three converging requirements
Three timelines are arriving simultaneously for European finance teams. FacturX mandatory compliance for large French companies creates a hard format-readiness deadline in January 2026. DORA — the EU Digital Operational Resilience Act binding since January 2025 — requires documented completeness of financial technology systems, including evidence that document intelligence captures all source types. European mid-market companies plan to increase supplier counts by 20% or more over the next 18 months (Euler Hermes, 2025) — each new supplier adding new format variants, new invoice layouts, and new configurations to an already diverse mix.
Companies without format-agnostic intelligence face a compounding problem: more suppliers means more format variants means more manual reconciliation hours per week. The scale tipping point arrives faster than most finance teams anticipate. Any organization with more than 50 active suppliers needs format indifference, not a future where suppliers standardize. The former is achievable now. The latter has been "coming" for thirty years.
Once every invoice, purchase order, delivery confirmation, and contract amendment — in every format — feeds the same index, a second-order capability emerges: supplier intelligence that was previously invisible. Supplier X raised prices on specific SKUs by 3% per quarter for six consecutive quarters. That pattern is visible only when 12 document types across 4 formats are indexed together. A query that previously required a multi-system data project becomes a 14-second question.
Ask your current AP tool today: does it query PDF invoices and EDI invoices in the same search? The answer reveals immediately whether you have format intelligence or format gaps. In 2026, your auditors will be asking the same question.