Where we started

In most organisations, supplier invoices are paid on trust in the contract. A controller rereads a few of them each month (the largest, the strategic suppliers) and the rest goes through. That is not negligence: available time decides what gets looked at, and sampling becomes the norm.

The problem is what that sampling lets through. A unit price a few cents above the negotiated rate, a volume rebate that stops applying after a reference change, a revision calculated on the wrong index. None of these discrepancies triggers an alert, because nothing is looking at them. And a share of spend (off-catalogue orders, recurring services, subcontracting) escapes the standard circuits entirely.

We founded Zylio in August 2024 to answer one precise question: how do you check every line of every invoice against what was actually negotiated, without adding headcount or replacing the ERP?

What Zylio does

Zylio is a platform of AI agents that reads your purchasing documents (framework contracts, price schedules, quotes, purchase orders, invoices) and confronts each invoiced line with the clause that governs it. When a discrepancy appears, the agent shows it: the contract clause and the invoice line highlighted side by side, with the amount at stake.

Concretely, three agents work together:

  • Capture extracts data from your documents, whatever their format, and keeps every value linked to the exact place it was read.
  • Compare puts quotes and supplier offers on a common basis, so the comparison bears on equivalent lines.
  • Compliance checks that what is invoiced respects what was negotiated: prices, rebates, revisions, quantities.

Nothing is committed without human approval. Zylio prepares the case; you decide. We do not touch the ERP, we contact no supplier, and no data is used to train a model. These three limits were set on day one and have not moved since.

Who it is for

Zylio is built for procurement leaders, finance leaders and accounts payable teams who handle a document volume nobody can reread by hand: dozens of active suppliers, framework contracts running to thousands of price lines, invoices arriving in every format. The private sector and the public sector alike: the unit price schedules of public contracts pose exactly the same problem as the framework contracts of a hotel group.

The platform integrates with existing ERPs, SAP, Sage, Oracle, Microsoft Dynamics 365 or NetSuite, without modifying them: Zylio reads what comes out, and returns its conclusions where your teams already work.

What we do not promise

We do not promise “automatic” margin or savings promised in advance. What we show is what has already been paid outside your agreements, line by line, with proof. What you do with it (recover a credit note, renegotiate, change supplier) remains your decision. On the scopes we have analysed, the recoverable gap sits between 1 and 7% of the spend analysed; the top of the range is never an expected result, it is an observed bound.

The team

Zylio was founded by four co-founders who come from enterprise SaaS, cloud and product: twenty years of enterprise sales, seven years at AWS on finance accounts, LLMs deployed at scale. This dual competence, procurement and artificial intelligence, is why the platform is designed for finance departments, not for laboratories.

This blog

We will write about four subjects: supplier spend control, agentic AI applied to procurement, the compliance of invoices with negotiated agreements, and the daily work of accounts payable teams. No figure your data would not confirm, no competitor named, no needless jargon.

Welcome to Zylio. If one of these lines sounds like you, the shortest route is to call us.