Rwanda consortium wins the AI-driven fraud and scams challenge at the C:>DIR Global 'Agentic Regulator' Hackathon
The National Bank of Rwanda (BNR) and the Rwanda consortium won the AI-Driven Fraud & Scams challenge at the first C:>DIR Global 'Agentic Regulator' Hackathon. Their AI agents help scam victims report fraud and recover their funds. We were proud to be part of the consortium, alongside Proto, the Global Legal Entity Identifier Foundation (GLEIF), and The University of Manchester. The finalists pitched at the C:>DIR Summit in Cambridge on 18 September 2026.
Cambridge Digital Innovation for Regulation (C:>DIR) and Financial Innovation for Impact ran the hackathon to test how agentic AI could help regulators and supervisors do their jobs. Agentic AI means AI agents that can carry out a task by themselves. The competition drew 942 builders from more than 65 countries, who submitted 336 concept notes. Of these, 36 finalists built working solutions in one week. Then 260 regulators and policymakers from more than 40 countries chose the final six.
From complaints to scams in one conversation
BNR already runs Intumwa, an AI agent that lets consumers file financial complaints in Kinyarwanda, English, and French. Average resolution time has fallen from 35 days to four. For the hackathon, the consortium extended Intumwa from complaints to scams:
A victim reports a scam in conversation.
The case is structured in about a minute.
FNA's Money Trails traces the stolen funds across mule accounts in real time.
People still make the final calls. Freezes, reimbursements, and cross-border requests need approval from a named person at an authorised institution.
The speed of reporting decides how much can be recovered
Instant payments move money in seconds, so stolen funds often pass through several institutions before the scam is reported. Fast reporting only recovers funds when every institution the money passes through can act on the report together. In Malaysia, FNA's National Fraud Portal raised the share of funds frozen in promptly reported cases from 0.5% to 21.5%. Bank Negara Malaysia reports a 75% reduction in time-to-trace.
Gillian Tett wrote about the hackathon in her Financial Times article, "How can we tackle AI's financial risks?", on 25 September 2026.
Thank you to Financial Innovation for Impact, Cambridge Digital Innovation for Regulation, and the C:>DIR Hackathon Team for running the event. Congratulations also to the other winning teams. All finalists and demos are on the C:>DIR winners page, and the full announcement is on the Financial Innovation for Impact website.
Learn more about how FNA helps central banks, payment system operators, and financial intelligence units tackle fraud and scams: fna.fi/products/fraud-portals
