AI Invoice Data Extraction Software
Data Alchemy reads your supplier and customer invoices — PDFs, scans, XML and FatturaPA — and posts them into SAP, Zucchetti or TeamSystem with no re-keying: your finance team only reviews the exceptions.
What invoice data extraction is
Invoice data extraction is the process that turns an invoice — in any format — into structured data ready for accounting: supplier, document number and date, taxable amounts, VAT, due dates and every single line item. Done by hand it is slow and error-prone; with traditional OCR templates every new supplier needs a configuration. Data Alchemy assigns a single AI engine to each invoice model — Claude by Anthropic (the engine we recommend by default), GPT by OpenAI, Data Alchemy AI (Data Alchemy's proprietary model, running on its own servers and GPUs in Italy) or another compatible large language model (LLM) through the European inference providers Scaleway and Seeweb Regolo.ai, on the customer's own subscription — understands the document semantics and extracts the fields with no template at all. The extracted data is validated in real time against the ERP master data and posted automatically: finance stops transcribing invoices and reviews only genuine exceptions. Automating accounts payable means extending this control across the whole flow, from purchase order to payment.
What data it extracts from every invoice
From each invoice Data Alchemy extracts both the header data and every line item, already normalised and ready for posting. Here are the typical fields recognised automatically.
Supplier data
- Company name and address
- VAT number and tax code
- Recipient code / certified email
- IBAN and payment terms
Header data
- Invoice number and date
- Net amount, VAT and document total
- VAT rates and nature (exemptions, reverse charge)
- Due dates and payment methods
Line items
- Item code and description
- Quantity and unit of measure
- Unit price, discounts and line amount
- References to order and delivery note
Tax data and controls
- Withholding tax and stamp duty
- Tax codes and social-security fund
- Totals and rate reconciliation
- Duplicate check against history
Invoice data extraction in 5 steps
From invoice receipt to accounting posting, Data Alchemy automates every step of the data extraction with no manual intervention.
1. Acquisition
The invoice arrives via email (from business mailboxes, through the Mail service installed at the customer's premises), scanner or upload. Only emails matching a model reach the platform.
2. Classification
The AI recognises that the document is an invoice — PDF, scan, XML or FatturaPA — and identifies its supplier, with no predefined templates required.
3. Data extraction
A dedicated LLM reads the header and line items and extracts every field: supplier, amounts, VAT, due dates and each line with item code, quantity and price.
4. Validation
The extracted data is checked in real time against ERP master data: suppliers, items and price lists are verified, totals reconciled and duplicates flagged.
5. Posting
The structured, validated invoice is written into SAP, Zucchetti or TeamSystem through the AI agent or Data Alchemy's REST API and webhooks, ready for accounts payable.
Everything you need to extract invoice data
Template-free AI extraction
A dedicated LLM reads invoices from any supplier and layout and extracts header and line items with no template configuration, at 99.8% accuracy (measured with Claude).
PDF, scans, XML and FatturaPA
Extract data from invoices in any format: native and scanned PDFs, images, XML and FatturaPA. One engine for all your invoicing flows.
Line-by-line reading
Beyond the header, every line item is extracted with item code, quantity, unit of measure, price and discounts, ready for matching against orders and delivery notes.
ERP master-data validation
Extracted data is validated in real time against SAP, Zucchetti or TeamSystem: suppliers, items, price lists and payment terms always consistent.
Duplicate and anomaly detection
Every invoice is compared with the posted history: duplicates, double submissions and tampered totals are blocked before posting.
Licence and pay-as-you-go credits
A Basic, Business or Enterprise licence chosen to fit your needs, plus pay-as-you-go credits for the documents processed. Quote on request.
The ROI of automated invoice extraction
Automating invoice data extraction frees finance from manual transcription and shortens posting times. Here are Data Alchemy's reference figures, the same as on the home page.
invoice data extraction accuracy with Claude, versus the errors of manual data entry
Per 1-page invoice (Claude Sonnet)
operational time (Data Alchemy average)
templates to configure: the AI reads any supplier with no setup
Solutions related to invoice extraction
Supplier invoice automation
The solution built for the finance team: post invoices and entries with no re-keying, validated against the ERP.
Learn more →Accounts payableAccounts payable automation
From order to invoice: data extraction, three-way matching and automatic posting into the ERP.
Learn more →ReconciliationInvoice reconciliation & three-way matching
Invoice-order-delivery note matching line by line with discrepancy detection before posting.
Learn more →ComparisonBest data extraction software
Invoice data extraction software compared by usage profile, with pricing models, Italian accounts payable criteria and sources.
Learn more →PlatformAll IDP solutions
The IDP (Intelligent Document Processing: automated reading of documents with AI) platform that brings every module together: invoices, delivery notes, orders, contracts and price lists in one flow into your ERP.
Learn more →Frequently asked questions about invoice data extraction
What is invoice data extraction?
Invoice data extraction is the process that turns an invoice — PDF, scan, XML or FatturaPA — into structured data ready for accounting: supplier, number and date, taxable amounts, VAT, due dates and every line item. Data Alchemy uses a dedicated LLM for each invoice model — Claude by Anthropic, GPT by OpenAI, Data Alchemy AI (Data Alchemy's proprietary model) or a compatible model through the European providers Scaleway and Regolo.ai — and, with Claude, reaches 99.8% accuracy (no accuracy figure is published for the other engines) and, with Claude Sonnet or higher, around 12 seconds for a one-page document and around 18 seconds for two pages (average times), with no templates and no manual re-keying.
Which invoice formats can it read?
Data Alchemy extracts data from invoices in any format: native and scanned PDFs, images, XML invoices and FatturaPA. It ingests documents from email (business mailboxes, via the on-premise Mail service), scanners or upload, classifies them automatically and extracts header and line items with no per-supplier template.
What data does it extract from an invoice?
From each invoice it extracts both the header data — supplier, VAT number, document number and date, net amount, VAT, total and due dates — and every line item with item code, description, quantity, unit of measure, unit price, discounts and amount. It also recognises tax data such as withholdings, stamp duty, tax codes and VAT nature, and handles references to orders and delivery notes.
How does it integrate with the ERP (SAP, Zucchetti, TeamSystem)?
Data extracted from invoices flows directly into SAP, Zucchetti, TeamSystem and other ERPs through the AI agent installed at your premises (a database, CSV/JSON/TXT files or APIs) or with an integration built on Data Alchemy's REST API and webhooks. Master data (suppliers, items, price lists, payment terms) is read from the ERP in real time, so the AI validates every invoice against the system of record before posting.
How is it different from traditional OCR?
Traditional OCR converts pixels into text but does not understand the invoice structure and needs a template for each supplier. An LLM-powered engine understands the document semantics: it distinguishes header from line items, recognises suppliers, items, units of measure and VAT rates, validates data against ERP master data and produces a structured invoice ready to post — with no per-supplier setup.
How much does automated invoice data extraction cost?
The cost of Data Alchemy is made up of a licence — Basic, Business or Enterprise, chosen to fit your needs — plus pay-as-you-go credits for the documents processed. From the Business licence up, external fields (the document checked against the ERP data it refers to, such as the purchase order) and remote validation (individual values, such as the supplier or item code, checked against your master data) check the AI-read data by querying your ERP via API. Quotes are on request and the licences are compared on the pricing page. Use the ROI calculator to estimate the savings on your real volumes.
See your invoice data extraction live
Activate the free Demo with 10 documents and check the AI reading on your real invoices, with 1 model chosen from the predefined ones in a dedicated environment. To see validation against your ERP and calculate the ROI for your case, contact us to plan a call with a demo on your own documents.