Procuros wants businesses to move beyond EDI, yet its own infrastructure still speaks EDI/EDIFACT, X12, XML, CSV, Excel, JSON, PDFs, APIs, AS2, SFTP and supplier portals. That apparent contradiction explains the company better than the phrase “autonomous trade.” Decades-old formats will not disappear simply because AI agents arrive. Procuros is trying to make those differences disappear from the workflow, translating them into structured transaction data before either an employee or an agent has to deal with them.
That thesis now has another €20 million behind it. Hamburg-based Procuros has raised a Series A led by XAnge, with W23 joining the round and existing investors Point Nine, Creandum and b2venture participating. The company serves more than 300 customers across DACH and the UK, and says over €3 billion in annualized order volume already moves through its network. The capital will support expansion across Europe and the United States, a larger go-to-market operation and further development of its AI-native supply-chain infrastructure.
The distinction between moving data and understanding a transaction is where Procuros is making its larger AI argument. Traditional OCR can pull text from an invoice or purchase order. Procuros says its platform additionally checks transaction data against master data, contracts and order history, matching identifiers such as GTINs and buyer or supplier IDs while validating quantities, prices, tax, totals, delivery information and payment terms before the data reaches an ERP. The company positions that accumulated context as the difference between an AI agent merely reading a document and safely deciding what should happen next.
That becomes clearer when the platform is viewed from both sides of a transaction. Suppliers can use Procuros for EDI connectivity, AI-based order ingestion, logistics automation, e-invoicing, payment reconciliation and SSCC/NVE labeling. Retailers get a different set of workflows around supplier connectivity, data validation, e-invoicing and three-way invoice reconciliation. Its retailer reconciliation product, for example, compares invoices against purchase orders and goods receipts and surfaces line-level discrepancies rather than treating the invoice as an isolated document. Procuros also connects with ERP environments including SAP, Microsoft Business Central, Odoo, Xentral and weclapp instead of requiring companies to replace the systems where their operations already run.
Company-published customer cases provide early evidence of how that abstraction can matter operationally. Luxury-beauty retailer Niche Beauty says it digitized more than 300 active suppliers through Procuros, including onboarding 100 suppliers in a single week, while reporting 90% less manual administration for its buying team. KoRo, meanwhile, uses the platform across more than 300 trading partners on both the supplier and retailer sides; its case study reports more than 20,000 invoices processed annually and a 90% reduction in manual processing time. Those are customer results published by Procuros rather than independent benchmarks, but they provide a more concrete test of the network model than the company’s broader automation claims.
The Quality Group offers another useful clue about where the product is headed. The owner of ESN and More Nutrition connected more than 30 retailers while moving its ERP environment to SAP, with Procuros reporting 95% less manual processing after the migration. That matters because supply-chain connectivity is rarely a clean greenfield project. Retailers change formats, suppliers have different technical capabilities, companies replace ERP systems and mandatory e-invoicing rules vary across markets. An autonomous-trade layer has to survive those changes without turning every new partner or system migration into another integration project.
Procuros has been working toward that abstraction since 2021, when its founders started with a narrower problem: businesses were still manually moving transactions between ERP systems, procurement platforms and supplier portals. The current product expands that idea from integration middleware into infrastructure for agents. Instead of asking an AI system to navigate the same fragmented environment humans struggle with, Procuros wants to normalize the environment first and expose reliable transaction context through its platform, APIs and MCP connectivity. CEO and co-founder Patrick Thelen summarizes the dependency simply: “AI agents are only as good as the infrastructure they run on.”
The involvement of W23 makes that ambition especially notable. Procuros is not only solving connectivity for software startups or individual procurement teams; its network sits between suppliers and retailers, where adding another participant can increase the usefulness of the infrastructure for everyone already connected. XAnge says the company has demonstrated the model in DACH and is now backing its attempt to carry it into larger international trading networks.
The €20 million Series A therefore sets up a more demanding test than simply adding customers. Procuros now has to show that the same transaction context can remain accurate across different ERPs, document formats, regulatory regimes and exceptions while AI agents assume more of the operational work. The intelligence may increasingly come from the agent. The harder infrastructure problem is making sure the purchase order, invoice or shipment it acts on means exactly what the agent thinks it means.
