Digital assets are changing more than the form in which value is held. They are reshaping when transactions occur, where business rules are applied, and how rapidly financial risk can spread. Tokenized deposits, stable coins and tokenized securities can settle conditionally, operate across platforms and continue outside traditional processing windows. The IMF describes tokenization as a structural change in financial architecture, one that can enable atomic settlement, continuous liquidity management and embedded compliance, but may also amplify instability through speed, concentration and fragmentation if governance does not keep pace.  

For banks, this creates a fundamental control question: Can a compliance model designed around downstream screening, batch monitoring and retrospective investigation govern transactions that execute through software in near real time? Increasingly, the answer is no. Compliance must remain independent and accountable, but its controls need to move closer to the transaction such that it operates before, during and after execution.

Why traditional compliance checkpoints fall short 

Conventional financial-crime controls typically operate across a sequence of systems: customer onboarding, sanctions screening, payment processing, transaction monitoring, case management and regulatory reporting. Programmable finance introduces additional surfaces such as wallets, smart contracts, distributed ledgers, custody providers and tokenization platforms, while linking on-chain activity to customer, account, treasury and risk data held off-chain.

The underlying regulatory obligations have not disappeared. The Financial Stability Board’s principle remains “same activity, same risk, same regulation.” Yet its 2025 review found significant gaps and inconsistencies in the implementation of crypto-asset and stablecoin frameworks, creating opportunities for regulatory arbitrage and complicating cross-border oversight. FATF’s July 2026 update similarly reported continuing challenges in translating virtual-asset risk assessments into effective controls, operationalizing licensing regimes and delivering effective risk-based supervision and enforcement. 

The implication is not that every compliance process must become instantaneous. Banks need to distinguish between controls that must operate in the transaction path and activities that can occur asynchronously. More ambiguous questions such as linked-wallet behavior, unusual transaction patterns, source-of-funds concerns or cross-jurisdiction conflicts which may require investigation and human judgment.

Building a Programmable Compliance Fabric

A more scalable model is a programmable compliance fabric with a policy, identity, monitoring, workflow and evidence services surrounding the transaction rather than a single control engine embedded inside one ledger.

  • At its foundation is policy-as-code. Approved policies are translated into versioned, testable rules governing who may participate, what assets may move, which jurisdictions are permitted and when a transaction must be held or escalated. 
  • Compliance and legal functions must remain responsible for policy meaning, approvals and change governance. Code makes approved policy consistently executable; it does not determine what regulation should mean.
  • Identity and permission controls then connect customers, counterparties, wallets, accounts and institutional roles.
  • AML and know-your-transaction (KYT) monitoring combines conventional customer information with wallet risk, transaction history, asset type, counterparties and on-chain behavior.
  • Travel Rule workflows validate originator and beneficiary information, identify missing data and retain the evidence exchanged with counterparties.

The control model described below combines these established obligations with emerging programmable capabilities.

The Role of Agentic AI in Programmable Compliance Fabric

Agentic AI should not become an autonomous compliance officer. Its defensible role is narrower: monitor, correlate, triage, gather evidence, orchestrate workflows and escalate exceptions.
An agent may bring together a transaction hash, customer profile, wallet history, custody record, applicable policy and previous case material. It can summarize why an alert fired, identify missing evidence, prepare an investigation narrative and route the case to the right reviewer. A separate agent might correlate an on-chain settlement failure with a custody status, ledger posting and treasury event to suggest a likely root cause.

Clear decisions should still be handled through deterministic rules. Material actions such as freezing funds, rejecting a transaction, overriding policy, filing a regulatory report or releasing an exception should remain subject to segregation of duties and authorized human approval. The architectural principle to be followed is to leverage rules and analytics for detection, AI for evidence synthesis and decision support, and humans for accountable decisions.

Architecture matters more than the ledger choice

Banks are unlikely to operate a single digital-asset platform. Their target state must connect distributed ledgers with custody platforms, payment rails, core banking, customer identity, KYC, transaction monitoring, treasury, case management and regulatory reporting. Therefore, a technology-agnostic integration layer is required with event adapters, reusable APIs and a canonical data model spanning customers, wallets, assets, transactions, alerts, cases, policies and evidence. The deterministic decision path should remain separate from asynchronous investigation. This allows a time-sensitive transaction to receive a policy response without waiting for a complex AI workflow, while exceptions receive deeper investigation afterward.

At the Wipro Innovation Network, we have enabled banks to implement digital assets ecosystem monitoring and compliance solutions. Building on this experience, we are exploring an advanced-forensics layer using Agentic AI to reduce manual investigation and improve handling of false positives. Our provider-agnostic digital-asset operations and compliance fabric follows this pattern: 

  • Pre-built adapters and configurable mappings connect blockchain, custody, core-banking, KYT and case-management systems
  • Rules support low-latency decisions; agents investigate exceptions
  • An evidence layer retains rule versions, recommendations, approvals and transaction lineage.

The objective is not to replace existing bank controls, but to orchestrate and integrate with them consistently across new asset classes and rails. 

What Banks Should Design Now
CXOs and compliance leaders do not need to predict which token or ledger will dominate. They do need capabilities that can evolve as products, infrastructure and regulation change:

  • A common control taxonomy across tokenized deposits, stablecoins, securities and custody activities.
  • Versioned policy services with maker-checker approval, testing and controlled deployment.
  • Real-time identity, wallet and transaction context spanning on-chain and off-chain data.
  • Travel Rule and cross-jurisdiction workflows that manage data gaps and exception.
  • End-to-end evidence lineage from policy and transaction through alert, investigation and human decision.
  • Governed agentic assistance with restricted tool access, explainability, monitoring and explicit decision rights.
  • Interoperability by design, so controls are not locked to one blockchain, custody provider or compliance vendor.

The strategic opportunity is to make compliance continuous, executable and evidence-ready. Banks that build this control fabric now will be better positioned to participate in programmable finance which holistically analyzes both on the on-chain and off-chain world. In the next generation of financial infrastructure, compliance will not simply review the transaction. Increasingly, it will help shape how the transaction is allowed to happen.

About the Author

Hitarshi Buch
Chief Architect and Frontier Tech Innovation Lead

With over 25 years of IT experience, Hitarshi specializes in enterprise architecture and frontier technology themes such as Blockchain and Quantum Computing.