Chapter 33 · AML, KYC and the Virtual Asset Travel Rule
Every fraud of Chapter 14 and every mule account of this chapter asks the same question of the system: who let this account open, who watched it, and who was told? The answer is a machinery most customers never see — customer due diligence, transaction reporting clocks, and a rule that now follows identities across blockchains. This chapter teaches the machinery precisely, then hands you the designer's seat: because the gate that keeps criminals out is the same gate that Chapter 32's excluded wait at, and a well-designed gate is the difference between the two chapters' problems solved or traded.
🎯 Learning Objectives
- Explain the money-laundering cycle — placement, layering, integration — and the AML triad against it
- Apply KYC mechanics: customer due diligence, risk categorisation, e-KYC, video KYC, beneficial ownership, PEPs
- State the reporting duties: the ₹10-lakh CTR with the integrally-connected-series rule, the seven-working-day STR, five-year records
- Explain the virtual-asset travel rule: what travels, above what threshold, and why both ends must be licensed
- Read the mule-account war's evidence: nineteen lakh accounts, MuleHunter.AI, the I4C-RBIH alliance
- Design a risk-based KYC process with proportionate controls and named reporting triggers
1. The Gate
Laundering has three classic stages. Placement: crime's cash or stolen funds enter the financial system — a deposit, a wallet top-up, a mule account. Layering: the money is moved, split, hops accounts, currencies, instruments — including, now, chain-hopping across virtual assets — until its origin is obscure. Integration: the cleaned value returns as apparently legitimate wealth. Against this cycle, the Prevention of Money Laundering Act, 2002 (PMLA) and its Maintenance of Records Rules build a triad: know the customer before the relationship starts (KYC/CDD), keep the records for five years, and report the flagged flows to the Financial Intelligence Unit, India (FIU-IND) — the national node that turns lakhs of individual reports into intelligence for investigators.
Hold the design tension from the start, because this chapter sits deliberately beside the last one. Every KYC document demanded, every threshold monitored, every account frozen is a cost imposed on all customers to catch the few — and every cost at the gate lands heaviest on those with the least documentation, exactly Chapter 32's excluded. The gate excludes and protects simultaneously; the designer's job is not to maximise either, but to place it deliberately. Everything that follows — mechanics, clocks, the travel rule, the mule war — is the material you place it with.
2. KYC Mechanics
Customer due diligence has four questions at its core, and every KYC innovation is a new way of answering them cheaply at a billion-person scale. Who are you? — identity verified against official documents or Aadhaar. Who really owns this? — beneficial ownership behind companies and trusts, at prescribed thresholds, so shells cannot hide the human. Why are you here? — purpose and nature of the relationship, the question that makes later behaviour interpretable. How risky are you? — customers are categorised low, medium or high, and the categorisation drives everything downstream: how much documentation, how much monitoring, how often the file is refreshed. High-risk categories carry enhanced due diligence — including politically exposed persons, whose prominent public functions earn extra scrutiny.
India's distinctive achievement (Chapter 4's story, now in full) is that these questions can be answered at near-zero marginal cost: e-KYC via Aadhaar OTP for instant basic onboarding; video KYC (V-CIP) — a live, recorded, geo-tagged video verification — for full accounts remotely; small accounts as a documentation-light formal entry; and the Central KYC Registry so a customer verified once is verified everywhere, ending re-submission of the same documents bank after bank. Periodic updation refreshes the file on a risk-based cycle. The inclusion dial is visible right inside the mechanics: minimum-KYC wallets work at small balances with restricted features, and full privileges unlock only with full verification — proportionality written into product architecture.
3. The Reporting Duties
The clocks, thresholds and retention — the compliance skeleton every reporting entity runs on, under the PMLA Rules and FIU-IND instructions.
| Duty | Trigger | Clock / form |
|---|---|---|
| Cash Transaction Report (CTR) | All cash transactions above ₹10 lakh — monthly, by credit and debit summation computed separately | Monthly; individual sub-₹50,000 items need not be detailed |
| Integrally connected series | Cash transactions individually below ₹10 lakh, aggregating above it within a month | The structuring trap — reported with the CTR |
| Suspicious Transaction Report (STR) | Proceeds of crime suspected; unusual or unjustified complexity; no economic rationale — cash or not | Principal Officer files within seven working days of forming suspicion; tipping off prohibited |
| Other prescribed records | Non-profit receipts above ₹10 lakh; counterfeit-currency transactions | Reported to FIU-IND per rules |
| Record retention | Transaction records, CDD files, EDD findings | Five years from the transaction; five years after the relationship ends, whichever applies |
| VDA service providers | Under PMLA since March 2023; consolidated FIU guidelines (January 2026) | 54 registered SPs; Principal Officers based in India; transfers treated on wire-transfer lines |
Two entries deserve their sentences. The integrally-connected-series rule is the law's answer to structuring — the deliberate splitting of a large cash movement into small pieces under the threshold; a customer depositing ₹9.5 lakh in pieces across a month is more visible, not less. And the STR's seven-working-day clock runs from forming suspicion — not from proving anything — with an absolute prohibition on tipping off the customer, which is why an STR never appears in your bank statement and why staff who know are legally silent. The VDA row extends the entire table to virtual-asset service providers: same duties, same FIU, new assets.
4. The Travel Rule
A wire transfer has always carried a passenger: the originator's and beneficiary's identifying information, moving with the money from institution to institution, so any investigator who grabs one leg of the trail can read the whole. The FATF's travel rule extends that passenger to virtual-asset transfers: above prescribed thresholds — commonly the equivalent of a thousand dollars cross-border — the originating VASP must transmit the originator's name and identifying details to the beneficiary VASP immediately and securely. India's FIU-IND guidance does this by treating VDA transfers on the lines of wire transfers, binding its 54 registered service providers to the same logic.
The rule's structural requirement is the point most commentary misses: it needs a licensed intermediary at both ends. A transfer from a registered Indian exchange to a registered exchange abroad carries the passenger; a transfer to a self-hosted wallet in a jurisdiction that has not implemented the rule carries nothing — the information has nowhere to arrive. This is the FATF's own scoreboard problem: 83 per cent of surveyed jurisdictions have travel-rule legislation on the books (up from 73 per cent in a year), but the named gaps are precisely the unhosted wallets, the offshore VASPs beyond supervision, and DeFi arrangements no licence reaches — the "sunrise problem," where the rule works only when dawn has reached both hemispheres. Chapter 29 asked which policy fork gives the rule the most reach; this chapter shows the mechanism the question is about.
5. The Mule War
The laundering cycle's Indian present tense is the mule account: an ordinary account — often bought or rented from a low-income owner for a few thousand rupees, sometimes sourced by compromised bank staff — that receives fraud proceeds and forwards them in a spidering pattern until the trail frays. The scale is now public arithmetic: the government told Parliament's committee of over nineteen lakh mule accounts identified and ₹2,038 crore of suspicious transactions prevented; the CBI's Operation Chakra-V searched 42 locations, arrested nine, and mapped about 8.5 lakh mule accounts across 700-plus branches; and the geography keeps moving — from Jamtara and Nuh to quieter districts in Madhya Pradesh, Karnataka and Assam, with laundering routes running through innovations as mundane as in-game purchases.
The counter-weapon is Indian fintech turning on itself: MuleHunter.AI, built by the Reserve Bank Innovation Hub — an AI model reading nineteen behavioural signatures across the banking system, operational in more than two dozen banks, detecting about twenty thousand mule accounts a month at 85–90 per cent claimed accuracy. In May 2026 the Home Ministry's cyber coordination centre (I4C) signed an MoU with RBIH to feed its Suspect Registry into MuleHunter's models — the police's list training the banks' detector. Draft RBI procedure now standardises the response: a flagged transaction of even ₹1,000 can trigger a hold, escalating to the whole suspected-mule account, with time limits so freezes do not become punishment without process. Notice the loop closing across this book: UPI made movement instant (Part II), fraud industrialised the movement (Chapter 14), and the AML machinery of this chapter is where the system fights back — with the same rails, the same data, and Chapter 31's AI-governance questions attached.
6. Designing the Process
The session's deliverable. A risk-based KYC process is a designed object, and this is the design sequence:
- Segment the customer base. Personas, channels, products — a SHG saver, a salaried urbanite, a crypto trader and a cross-border merchant are different risks.
- Assign risk factors. Geography, product, channel, customer type, expected activity — the inputs to low/medium/high categorisation.
- Proportion the gate. Documentation and verification scale with risk: OTP e-KYC for small balances; V-CIP for full accounts; enhanced due diligence for high-risk, PEPs, complex ownership.
- Build the expected-activity baseline. The CDD answers ("why are you here?") become the yardstick — monitoring flags deviation, not absolutes.
- Wire the triggers. CTR thresholds with the integrally-connected-series logic; STR red flags routed to the Principal Officer with the seven-day clock; mule-pattern signals (Chapter 14's stack) into the same queue.
- Audit and iterate. Sample files, test the model, measure false positives — an AML system that flags everyone protects no one and excludes many.
And the gate's placement is a judgement to state, not hide: tighter gates catch more mules and exclude more first-time users; looser gates include more of both. The design answer is tiering — cheap entry at small exposure, escalating diligence at thresholds — which is why the minimum-KYC wallet and the travel rule live in one chapter. Chapter 34 asks the question this design begs: when a platform embeds a bank's products, whose gate is it, and whose regulator stands behind it?
🛠 Hands-On: KYC Process Design
Design the AML process for PaySakhi, a payments-and-credit app for women SHG members: savings wallets, UPI, small working-capital loans, and a new feature letting users buy ₹5,000 of crypto as an investment.
Then judge three accounts and one transfer:
- Account A: ₹9.5 lakh deposited in cash across six visits in one month, each below ₹2 lakh.
- Account B: receives ₹40,000, splits it into 40 UPI sends of ₹1,000 to strangers, then buys crypto with the remaining ₹8,000 — pattern repeats weekly.
- Account C: dormant for two years, wakes on a new device and new SIM, immediately receives ₹2 lakh from three unknown senders.
- The transfer: a user sends ₹90,000 in crypto from her PaySakhi account to a self-hosted wallet address in a jurisdiction with no travel-rule law.
Steps
- Build PaySakhi's process through the six design steps: segments, risk factors, tiered gates, baselines, triggers, audit — in table form. The inclusion constraint from Chapter 32 is a hard requirement: no document may be demanded that a typical SHG member cannot produce.
- For each account A-C: is it structuring, layering, or account takeover? Which report fires (CTR, STR, both, neither), and by when? Name the specific red flags.
- The transfer: what does the travel rule require PaySakhi to do — and what can it actually do, given the destination? Design the control you would implement.
- The counter-case (150 words): a reviewer says your design is too loose and PaySakhi will become a mule farm. State their strongest point, your defence, and the one change you would make.
View Solution / Walkthrough
1. Model process table. Segments: SHG saver (core), UPI-active trader, loan customer, crypto-curious investor. Risk factors: channel (remote onboarding), product (credit, then VDA), activity profile, cash intensity of the local economy. Tiered gate: Tier 1 — Aadhaar OTP e-KYC, wallet capped at ₹10,000 balance, UPI within caps; Tier 2 — V-CIP at higher limits or first loan, income conversation (not documents) via SHG group-endorsement; Tier 3 — enhanced due diligence before the crypto feature unlocks: full V-CIP, source-of-funds discussion, purpose noted. Baseline: expected monthly inflows ₹5-15k (SHG rhythm), deviation-flagged monitoring. Triggers: CTR engine with integrally-connected-series logic; STR red flags — velocity in-out, fan-out to unknown payees, dormancy + device/SIM change + inflow spike, VDA purchase immediately after fan-out — routed to the PO's seven-day queue; MuleHunter-style pattern sharing if the app joins a consortium. Audit: quarterly false-positive review, model fairness check (Chapter 31's sutras apply to fraud models too), board reporting.
2. The three accounts. A — structuring: six cash deposits individually far below ₹10 lakh, monthly aggregate above it: the integrally-connected-series rule catches what a per-transaction threshold would miss — a CTR fires (with sub-₹50,000 items aggregable), and the pattern itself is an STR red flag: no SHG baseline explains ₹9.5 lakh of cash; if the PO forms suspicion, the STR files within seven working days, no tipping off. B — layering: fan-out to 40 strangers then VDA purchase is the mule signature (in-out velocity, unknown payees, crypto exit); STR on suspicion of layering — the CTR likely never triggers (no cash), which is precisely why the STR exists; consider a transaction hold under the draft SOP's ₹1,000 flag logic. C — account takeover: dormancy + new device + new SIM + sudden multi-sender inflow is Chapter 14's takeover stack wearing AML clothes; this is fraud-first (freeze and verify, Chapter 14 machinery), then STR if the inflows prove to be scam proceeds routed to a fresh mule. Both reports can be right simultaneously.
3. The transfer. The rule requires originator information to travel with the transfer — but only to a beneficiary institution. To a self-hosted wallet in a no-travel-rule jurisdiction, there is no arrival point: the passenger cannot disembark. What PaySakhi can do: apply its own Tier-3 due diligence before permitting VDA withdrawals (the rule binds the regulated end), record the destination address, screen it against known-risk lists, set withdrawal caps, and treat repeated transfers to non-implementing jurisdictions as an STR red flag — the sunrise problem managed at the only end the sun has reached.
4. Model counter-case response. "The reviewer's strongest point is real: OTP-based Tier-1 is exactly what mule recruiters farm, and ₹10,000 caps multiply across dozens of bought accounts into serious capacity. My defence: the cap bounds each account's harm, the baseline monitoring catches mule rhythm (in-out velocity breaks the SHG pattern within days), and tiering is the only design that meets the no-document constraint — a documents-first gate would simply push the SHG base out of the formal system, which is how the predatory lenders of Chapter 16 got their market. The change I would make: velocity caps on Tier-1 UPI fan-out to unknown payees in the first ninety days — blunt, slightly annoying, and targeted precisely at the farming pattern — reviewed quarterly against false-positive data."
Key Takeaways
The AML triad — know the customer, keep the records five years, report the flagged flows — attacks laundering at placement, layering and integration.
The clocks: CTR above ₹10 lakh including integrally connected series; STR within seven working days of suspicion, never tipped off.
The travel rule makes identity travel with virtual-asset transfers — but only between licensed ends; self-hosted wallets in lagging jurisdictions are the sunrise problem.
Nineteen lakh mule accounts, ₹2,038 crore blocked, and MuleHunter.AI detecting 20,000 mules a month: the laundering war is now fought with the fraud stack itself.
A risk-based process is a designed object: segment, factor, tier the gate, baseline the behaviour, wire the triggers, audit — with the inclusion cost stated, not hidden.
Test Your Understanding
Choose one best answer per question. The reporting table and the hands-on's four cases are fair data. Aim for 24/30 before moving to Chapter 34.
1. The three classic stages of money laundering:
2. The AML triad built by PMLA and its rules:
3. Beneficial ownership exists to answer:
4. PEPs earn enhanced due diligence because:
5. Video KYC (V-CIP) enables:
6. The Central KYC Registry's core function:
7. The Cash Transaction Report fires on cash above:
8. The integrally-connected-series rule exists to catch:
9. The STR clock, correctly stated:
10. An STR never appears in your bank statement because:
11. PMLA record retention:
12. The travel rule's common cross-border threshold:
13. The travel rule's structural requirement:
14. FATF's Seventh Update scoreboard, and its gaps:
15. India's travel-rule implementation route:
16. A mule account, defined:
17. The scale numbers stated to Parliament:
18. MuleHunter.AI's claimed performance:
19. The I4C-RBIH MoU (May 2026) connects:
20. The mule geography has shifted from Jamtara and Nuh to:
21. Under draft RBI procedure, a suspected mule transaction flag starts at:
22. Account A's six sub-₹2-lakh cash deposits totalling ₹9.5 lakh in a month is:
23. Account B's fan-out-then-crypto pattern triggers:
24. Account C's signature reads as:
25. The sunrise problem, named:
26. The minimum-KYC wallet exemplifies:
27. A monitoring baseline exists so that flags fire on:
28. Why false positives matter to inclusion:
29. Chapter 34's question about the gate:
30. The chapter's closing rule of gate design: