A fraud score is not a defence. That, in a sentence, is the premise of the first challenge to run through VerityX Innovation Labs, the Emirates Institute of Finance-backed programme that aims to compress the UAE’s bank innovation cycle from twelve months to ninety days. The brief, AI-Powered Mule Account Detection and Interdiction for UAE Banks, asks fintechs for something the market has been slow to supply: a system that spots a money mule early and then does something about it, inside controls a bank can actually operate and defend to its regulator.
The target is a hard one precisely because it looks ordinary. Mule accounts are the settlement layer of almost every consumer fraud and laundering scheme running in the UAE. Victims of investment, romance, job and impersonation scams are directed to move funds into accounts controlled, knowingly or otherwise, by third parties, who then layer the money rapidly across banks, wallets and borders. The account itself is legitimate and KYC-verified, behaving in a subtly abnormal way, which is why rules-based monitoring routinely misses it until the money has gone. The brief notes that more than half of UAE residents report being targeted by a financial fraud attempt every month, with bank transfers and cards the most-targeted channels, and that mules are recruited continuously through fake job offers and social media, so the population of at-risk accounts refreshes faster than any blocklist can track it.
Regulation supplies the deadline. CBUAE Notice No. CBUAE/FCMCP/2025/3057 makes mule-account detection an explicit supervisory expectation, requiring licensed institutions to identify mule accounts early on the basis of unexpected or suspicious transaction patterns, to run real-time transaction monitoring around the clock with per-transaction risk scoring, and to stop or decline suspicious activity as it happens. Most of those requirements carried a compliance deadline of 31 March 2026, with penalties of up to AED 250,000 per violation. The supervisory expectation is therefore already live, and the gap between what banks are required to do and what their current stacks can do is what the challenge is designed to close.
The brief splits the problem in two. On detection, it asks entrants to surface accounts receiving and moving illicit funds before or as the money moves, using signals including rapid in-and-out flows, fan-in and fan-out patterns, pass-through balances that hover near zero, sudden divergence from an account’s own behavioural baseline across velocity, counterparties, channels, devices and geography, network links to already-suspected mules through shared devices and coordinated timing, and the tell-tale profile of newly recruited or dormant-then-active accounts showing signs of remote control.
The second half is where the brief departs from a conventional fraud tender. Detection must convert into what VerityX calls action to stop: a proportionate, timely intervention that halts loss without punishing legitimate customers. That means real-time hold, decline or step-up on outbound transactions from a suspected mule, prioritised and explainable alerts with the evidence behind each decision attached for fraud and AML investigators, a feedback loop so investigator outcomes retrain the model and drive false positives down, and support for inter-bank signal sharing where funds move across institutions, within legal and data-protection limits.
Entrants build and test inside the VerityX Labs Sandbox rather than against a slide deck. Participating fintechs get synthetic and anonymised UAE-representative transaction, account and device datasets seeded with mule typologies and labelled outcomes; an isolated environment for running detection models and interdiction logic against streaming and batch data without touching live customer systems; reference API patterns and event schemas that mirror how a sponsor bank’s core, payments and case-management systems would call the solution; a common evaluation harness with holdout scenarios so every entrant is scored on the same basis; and structured sessions with fraud, AML, data and regulatory specialists.
Scoring reflects the brief’s central argument that deployability is as decisive as accuracy. Detection performance, covering precision, recall, speed and the false-positive burden on genuine customers, carries 30 per cent. Action and interdiction, judged on the effectiveness, proportionality and safety of the stop mechanism and on measurable loss prevented, carries 25 per cent. Explainability and auditability, meaning clear per-decision reasoning a bank can defend to CBUAE and to its own customers, carries 20 per cent. Deployability, covering latency, scalability, integration effort and fit with existing bank controls, carries 15 per cent, and compliance and data ethics the remaining 10 per cent. A solution that detects brilliantly but cannot be run inside a UAE bank’s control environment will not win.
The programme is looking for fintechs and regtechs with working capability in fraud detection, transaction monitoring, behavioural analytics, graph and network analysis, or AI and machine learning for financial crime, able to deploy and iterate within the sandbox timeframe and to integrate by API. A UAE presence or licence is helpful but not required, and the brief explicitly welcomes early-stage teams with a strong, novel approach. The indicative timeline runs two weeks to register, one to accredit, two to build and test, and one to evaluate in front of sponsor banks and regulator observers, with selected solutions progressing to a controlled pilot with a sponsor bank.
That last phase is the point of the whole exercise. The Innovation Labs model exists to give the UAE’s banks a governed route from proof of value to production pilot without the multi-month procurement cycle that kills most bank-fintech pilots before they start, and the mule challenge is the first test of whether it works. Fintechs and regtechs wanting to take part can register at the VerityX Labs platform, referencing the Mule Account Detection Challenge, and complete accreditation to gain access.
The post Detect It, Then Stop It: Inside the First VerityX Labs Challenge for UAE Banks appeared first on The Fintech Times.
