Shared signals,
not shared data.
The MuleTrace mule network is designed to let institutions warn each other about confirmed mule accounts without sharing customer data. Members share one-way hashes of identifiers; a match raises the alert and shows a “confirmed mule at another institution” badge.
Why one bankcan’t see the ring.
A victim pays an account at a bank, the money hops to a finance company, then leaves through a payment app. Each institution sees only its own slice, so each slice looks small.
A member confirms a mule
After its own investigation, an institution marks an account as a confirmed mule.
Identifiers are hashed
NIC, phone, device and account numbers are normalised and hashed one way with SHA-256 and a network key, inside the institution.
Others get warned
When another member’s customer pays a matching account, the case jumps up the queue with a clear badge.
Head office never seesbank transactions.
No bank sends its transactions to the network. Head office only hears three things: a confirmed mule, where that mule’s money went, and scam reports from the public. That is enough to join mules at different banks into one chain.
A bank reports a mule and where the money went
Bank A confirms a mule with its own monitoring and reports its hashed ID, plus the bank and hashed account the money was sent to.
The receiving bank is told
Head office notifies Bank B that money from a confirmed mule arrived in one of its accounts. Bank B checks that account using its own records.
The chain joins up
When Bank B confirms its mule and reports the next hop, head office links the two. Hop by hop, the whole route through banks, the stock market and out of the country is rebuilt.
Victims can report the account they paid
A teenager pays Rs 16,000 for a phone that never comes. Today that money is gone and nobody connects it to the other victims. With MuleTrace, anyone can report the account they paid. It is hashed on arrival, matched against confirmed mules, and when several independent reports from different districts name the same account, the bank that holds it is asked to review it.
Three months of history, before go-live
In a secure environment inside each bank, MuleTrace runs over the last three months of transactions. Only hashed mule findings leave the bank, so the network starts with the mules that are already active.
Nothing is blocked automatically
One report never flags anyone. Each bank decides using its own records, and every STR is approved by an analyst before it goes to the FIU.
What does a matchlook like?
In the demo ring, four mules at three fictional institutions pass money to one aggregator. The aggregator’s hashed ID is already in the network, so the analyst sees the badge on the graph before writing a word.
- Red ringVictims → four mule accounts → one aggregator → cash-out AGT-007.
- Badge“Confirmed mule at another institution” on the aggregator, from a hashed-ID match.
Is customer datashared between banks?
What would be shared
- One-way hashes of a confirmed mule’s NIC number
- Hashes of phone number, device ID and account number
- The fact that a member confirmed the account as a mule
What never leaves the institution
- Names, addresses or any raw identifiers
- Balances, transactions or KYC documents
- Which customer of yours matched, or why
- Your analysts’ notes and decisions
The network:your questions.
Last updated 13 October 2026
Is customer data shared between banks?
No raw customer data is shared. The network is designed so members share only one-way hashes of identifiers belonging to confirmed mules. A hash cannot be turned back into the original ID number, phone number, device or account.
What exactly gets hashed?
National identity card (NIC) number, phone number, device identifier and account number of accounts an institution has confirmed as mules. Each is normalised and hashed with SHA-256 and a network key before it leaves the institution.
Is this approved under Sri Lankan law?
Not yet established. The sharing model has to be reviewed by lawyers against the Financial Transactions Reporting Act and the Personal Data Protection Act, 2022, and agreed with participating institutions, before any live use. The prototype uses a synthetic network table.
Does head office see bank transactions?
No. Banks send only a confirmed mule’s hashed ID and the next hop: the bank and hashed account the money went to. Head office notifies that bank, which checks the account against its own records. The chain is built from these reports alone.
Can members of the public report a scam?
That is the plan. A victim reports the account they paid, which is hashed on arrival. A single report never flags anyone. Several independent reports, or a match with a confirmed mule, ask the holding bank to review the account. A live service needs agreement with CBSL, the FIU and the police.
What does an analyst see when there is a match?
The case moves up the queue and shows a “confirmed mule at another institution” badge on the matching counterparty. It does not reveal which institution reported it or any of that institution’s customer details.
Why would institutions join?
Mule rings spread across banks, finance companies and payment apps, so each one sees only part of the picture. The more members confirm mules, the earlier everyone else is warned.
Run a piloton your own alerts.
We are looking for a small number of Sri Lankan institutions to try MuleTrace on their own historical alerts. Tell us about your team and we will set up a walkthrough.