Cato raises 6M euro for tender AI: your bid library is the asset

Milan-based Cato raised a €6 million seed round led by Keen Venture Partners to expand its AI platform for finding and winning public tenders. The tender AI pitch is narrow and unglamorous: read every published bid opportunity, work out which ones you can actually win, and pre-fill the paperwork. It is also one of the clearest examples we have seen of a small-company workflow that was always a data problem pretending to be a headcount problem.

What actually happened

Per Tech.eu, the round brings Cato's total to €7.6 million after a pre-seed led by Italian Founders Fund. Existing backers Vento, Heartfelt, Moonstone, BHeroes, Alecla7, and Nova Venture followed, alongside angels including the founders of Lexroom, Sibill, and Pillar.

The product does four things, in order:

• Monitors 27,000+ sources in real time and ranks the tenders a given company is positioned to win
• Extracts requirements and evaluation criteria out of the tender documents
• Flags inconsistencies between documents before they cause a disqualification
• Pre-fills administrative and technical sections from the company's own past materials, with the human still submitting

Within six months of launch it has processed thousands of tenders for Italian SMEs and larger firms in medical devices, construction, IT, and services, according to EU-Startups. The market context: Italy's public tenders were worth €309.7 billion in 2025, up 13.9%, per the country's anti-corruption authority ANAC. Cato cites EU-wide public procurement at roughly 14% of GDP across 250,000+ contracting authorities — a figure worth treating as a market-sizing slide rather than a verified statistic.

Why this matters for your business

Most US small firms will never touch an Italian tender. Copy the architecture anyway, because the same shape sits inside every RFP response, insurance renewal, grant application, vendor security questionnaire, and municipal bid packet you have ever suffered through.

The third feature is the real product. Discovery is a search problem and lots of people sell it. Pre-filling is a generation problem and everybody sells it. But flagging inconsistencies across documents before they disqualify you is the one that maps to money, because losing a bid on a formatting technicality after 30 hours of work is the failure mode operators actually experience. When you scope your own document automation, start at the disqualification check, not the draft.

Your past proposals are a structured dataset you have never structured. Every firm we work with has four years of submitted bids sitting in a shared drive as PDFs and DOCXs, named final_v3_REALFINAL. That corpus is the only defensible input in this whole workflow — the model is a commodity, your answers to "describe your QA process" are not. The build is: extract, tag by question type, store as retrievable chunks, and route new questions against it. That is a two-week project, not a platform.

Keep the submit button human. Cato's own framing keeps the customer in control of submission, which is the correct call for anything with a legal signature attached. Automate the reading and the assembly. Do not automate the commitment.

The pattern generalizes past bids. It is the same one we described for procure-to-pay agents: the value is in parsing documents you did not write against records you do own.

Sources: Tech.eu, EU-Startups.