HKMA Explores Agentic AI to Catch Crypto Financial Crime

Hong Kong Central Bank Looks to AI to Fight Financial Crime
Financial regulators around the globe are constantly seeking smarter ways to monitor digital asset flows and prevent illegal activities. Recently, the Hong Kong Monetary Authority (HKMA), which serves as the city's central banking institution, revealed that it is examining the use of agentic artificial intelligence (AI) to combat crypto-related financial crime.
Speaking on the new initiative, HKMA Executive Director Raymond Chan shared that AI technologies could play a key role in analyzing complex transaction patterns. Specifically, AI systems can evaluate both onchain transfers happening directly on public blockchains and offchain customer transactions taking place within traditional financial institutions to flag anomalies.
What is Agentic AI and How Does It Work?
To understand why this approach is notable, it helps to know what makes agentic AI different from standard software. Traditional automated monitoring tools follow rigid rules, flagging transactions only when strict preset thresholds are breached. In contrast, agentic AI operates with greater autonomy.
Think of agentic AI as a digital investigator capable of processing massive volumes of data around the clock. Instead of relying solely on fixed triggers, agentic AI systems can evaluate complex contexts, reason through multi-step transactions, and continuously adapt to evolving methods used by illicit actors.
Connecting Onchain and Offchain Data Points
One of the primary challenges in tracking financial crime is that illicit funds often move back and forth between decentralized crypto networks and traditional banking systems. Raymond Chan pointed out that agentic AI could help connect these two separate layers of data:
- Onchain Transfers: Data recorded publicly on blockchain networks, showing wallet addresses, transaction volumes, and movement between protocols.
- Offchain Transactions: Real-world banking activities, including fiat deposits, wire transfers, and identity verification data managed by traditional institutions.
By cross-referencing onchain movements with offchain bank records, AI models can detect subtle irregularities that legacy systems might miss. For example, if a bank account receives funds linked to a flagged crypto wallet, the system can quickly alert compliance teams for review.
Why This Regulatory Exploration Matters
Hong Kong has actively positioned itself as a global hub for regulated digital assets. Creating a secure environment with strong anti-money laundering controls is critical for maintaining market integrity and protecting participants.
Exploring tools like agentic AI reflects a broader movement among central banks to leverage modern technology to enhance regulatory compliance without slowing down legitimate financial activities.
Latest blog posts

Bitcoin Volume Surges to Highest Level Since February
Bitcoin daily trading volume hit multi-month highs following a drop to $55,100, triggering $562 million in liquidations as prices bounced back.

Anthropic IPO, Saylor Strategy, and Quantum Bitcoin Security
A friendly weekly recap covering MicroStrategy plans, Anthropic IPO impacts on Bitcoin ETFs, Hong Kong regulatory alerts, and quantum security research.

Fed Rate Hikes Could Extend Beyond December as Inflation Stays High
Federal Reserve Governor Lisa Cook signals that interest rates might stay higher for longer, impacting both traditional stocks and cryptocurrency markets.