Polfund and 10Clouds Financial Institutions: AI That Turns Public Tenders Into Qualified Leads
WARSAW, July 2026
10Clouds Financial Institutions and POLFUND Fundusz Poręczeń Kredytowych S.A., established in 2001 and one of Poland's largest macro-regional funds supporting SME finance in the west of the country, have completed a joint AI project. Polfund provides credit and bid-bond guarantees in partnership with major banks like Santander and EU institutions. The two products delivered during the engagement, SmartLead AI and AIConsole, are now live in the fund's production environment.
Project challenge: automating lead generation and counterparty verification in public procurement
The public procurement market generates thousands of sales signals every day, but financial institutions offering bid bonds and guarantees do not struggle with a lack of data. They struggle with too much of it, scattered across sources, and the risk of picking the wrong clients. Polfund needed to move from manually browsing tenders to a fully automated lead generation and verification process.
The fund's business model depends on reacting quickly and safely when a tender appears and a contractor needs a bid bond. At market scale, however, data is buried in PDFs and protocols, contact details are often missing, and every prospect requires time-consuming checks across public registries. Manual analysis delayed outreach, increased verification risk, and left the sales team selecting prospects instead of talking to clients.
The system removes that burden. It continuously screens thousands of new proceedings and discards those that do not require a bid bond or guarantee. For the rest, it extracts contractors, bid values, and registry data such as NIP, REGON, and KRS from tender documentation. Each lead then goes through automated verification of legal status, business age, and debt signals before it reaches the sales team.
Key AI capabilities
- Automated source monitoring: scans up to 5,000 new tenders daily from e-zamowienia.gov.pl and 500 from ted.europa.eu.
- Dedicated site maps (whitelist): AI agents trained on specific portal structures achieve 95% data-extraction accuracy on key platforms.
- Business filtering: automatically selects tenders by bid-bond or guarantee requirement and contract value, up to PLN 100M for performance guarantees.
- Bid-participation prediction: estimates the likelihood that a contractor will bid on a proceeding based on historical patterns between tender and bidder characteristics.
"We wanted to give Polfund's team back their time, so they spend it talking to the right clients instead of searching for them. The system already helps the fund grow its business, reaching the right contractors earlier and with more confidence," said Karol Stępień, CEO of 10Clouds Financial Institutions.
Business impact
- 100% automated research: analyzing approximately 3,000–5,000 proceedings daily, with no manual work needed.
- 300% faster processing: handling twice the request volume in the same time.
- High-quality lead base: automatically excluding indebted, suspended, or under-a-year-old companies.
- Precise targeting: identifying only the top three bidders who genuinely need bid bonds.
- Data enrichment: recovering hard-to-find contacts from PDFs and external sources.
The fund moved from a manual, reactive way of finding clients to a fully automated, AI-driven process. Lead research that once consumed hours of manual work is now handled end to end by the system. Counterparty checks that used to take a specialist's time now happen in seconds, and low-quality prospects, dormant, indebted or too new, are filtered out automatically. The sales team spends its time on conversations, not on selection.
Most importantly, the fund now reaches the right contractors earlier and with more confidence than manual work ever allowed.
Read the full story in our Polfund case study.