Global AML and Financial Crime TechSprint 2018

In May 2018, we held a 3-day TechSprint, focusing on how technology can be used to more effectively combat money laundering and financial crime.

The United Nations (UN) estimates that at least US$1.6 trillion is laundered through the global financial system each year. This is the proceeds of human, drug, arms and endangered wildlife trafficking, slavery, corruption, fraud and other crimes. It is also estimated that less than 1% of illicit financial flows are intercepted globally.

The purpose of the TechSprint was to investigate how new technologies and greater international collaboration could help to improve prevention and detection rates.

Megan Butler, executive director of supervision – investment, wholesale and specialist, highlighted how data and technology can help detect and disrupt criminal activity.

Read Megan's speech[2].

Christopher Woolard, executive director of strategy and competition, also spoke on the importance of greater international collaboration in fighting financial crime.

Read Christopher's speech[3].

TechSprints[4]

Developing solutions

Over the 3 days, 260 participants from 105 firms spanning 16 countries – together with regulators and law enforcement agencies from the United States, Europe, Middle East and Asia Pacific – worked in teams to develop solutions to various problem statements.

Problem statements

How could the detection and reporting of suspicious activity within capital markets be improved?
How can technology aid to look at transactions flows across jurisdictions when a centralised way of data sharing is not possible?
How can the quality of feedback provided to firms from suspicious activity reporting be improved without breaching data privacy and anti-money laundering/tipping-off restrictions?
How can financial crime patterns be efficiently and effectively identified and codified in such a form that the algorithms can be shared across borders or between institutions?
How can robotic process automation be applied to improve the efficiency and/or effectiveness of regulated institutions’ financial crime and money laundering detection and prevention activities?
How can natural language processing aid in the investigation, automation of production and analysis of suspicious transaction/activity reports?
How can one institution’s suspicions of fraud be recorded in such a way as to reduce the risk of laundering those financial gains at another institution while managing the risk of tipping off and protecting personal data?
How can appropriate sanctions, money laundering and fraud assessments and interventions be implemented in distributed ledger-backed payment systems?
How might cryptographic tokenisation be used to aid identification and tracing of units of value through the international financial system?
How new technologies, big data, etc can be leveraged to measure the incidence of financial crime/money laundering/terrorist financing etc - to aid firms and regulators/agencies in understanding the effectiveness and impact of compliance, control, prevention and prosecution activities.
How can Network Analytics be used to analyse trends and patterns across the network to give a meaningful network risk score?
How can Machine Learning and Behavioural Science be used to facilitate decision-making for operational support systems?
How might technology (machine learning and AI) assist in the more effective identification of authorised payment fraud? i.e. where consumers are tricked into sending money to a fraudster.
How meta-data around transactions (such as IP addresses, card authorisations addresses, unique identification details of mobile phones and laptop computers, browsing data) can be incorporated in the analytical process?
How can technology be used to facilitate efficient on-boarding of a new customer whilst at the same time ensuring a robust KYC process and giving the customer the ability to protect their own data?

The teams then demonstrated their solutions to a cross-industry judging panel and a sizeable audience of senior executives from regulated firms, technology providers, start-ups and academic institutions.

Some of the technologies and prototype solutions that were produced and showcased by the teams included:

A shared database of ‘bad actors’, secured and distributed using distributed ledger technology. The database would allow a financial institution to query whether a new customer had been rejected by another financial institution due to financial crime activities or concerns.
Natural language processing, topic modelling and text analytics to enhance financial crime-focused transaction monitoring solutions within financial institutions.
Graph/network analytics to more readily identify relationships between entities to aid in due diligence and ongoing monitoring of potentially suspicious entities and activities.

The teams also developed various methods to enable:

Greater sharing of crime typologies/patterns between institutions to aid detection and intervention capabilities.
Querying by a financial institution of the confidential/encrypted data of another financial institution using homomorphic encryption and/or zero-knowledge proof technologies. These technologies could enable financial institutions to verify certain types of information with each other, without compromising the security or confidentiality of the underlying data.
Centralisation of data from multiple institutions into a shared utility, with the data then being analysed for fraud, money laundering and sanctions monitoring purposes.

Awards

Prizes were awarded for the following categories. The winning teams received support to progress their solutions from Level 39, RegTech Associates and The Disruption House.

Fast award: solutions that could be implemented most quickly

The winner, Connected Picture, developed a machine learning system that uses the transaction and network features of known ‘bad actors’ to prevent authorised payment fraud.

Eureka award: the most original idea

The winner, Catch the Chameleon, developed blockchain technology to share questions and answers between nodes in the financial system to identify matches between different profiles for the same entity without sharing data.

Jump award: potential to provide the biggest jump forward

The winner, Red Flag, developed a natural language processing system to provide additional meaning and context to transactions.

Artificial intelligence systems search for suspicious phrases within the transaction notes and analyse how common it is for certain industries to be transacting with each other to determine a risk score for transactions.

The winner, Connected Picture, developed a machine learning system that uses the transaction and network features of known ‘bad actors’ to prevent authorised payment fraud.

The winner, Red Flag, developed a natural language processing system to provide additional meaning and context to transactions.

Observing proofs of concepts

We also acted as an observer for several consortia developing anti-money laundering related proofs of concept resulting from the TechSprint.

These include:

Catch the Chameleon: a collaboration between Santander and other financial and technology industry partners using blockchain to allow the cross-referencing of KYC/ KYB data on small and medium-sized enterprises (SMEs) with suspicious characteristics.
Webs of Suspicion: a collaboration between Solidatus, Grant Thornton UK LLP, Ashurst LLP, Solace and Privitar, to explore ways to help firms safely detect and share data on financial crime, using machine learning AI, encryption and smart contract protocols, underpinned by open standards.
Pooled Transactional Data: a collaboration between Deloitte and 3 large UK retail banks to assess whether using advanced analytics techniques on pooled transactional data from multiple financial institutions may identify potential money laundering activity which would not have been identified by conventional practices.

Find out more

A podcast recorded at the TechSprint with Christopher Woolard and Nick Cook[5], provides more detail on the background and ambitions of the TechSprint.

For more information, email [email protected].

First published: 31/05/2018 Last updated: 18/12/2025

Global AML and Financial Crime TechSprint 2018

Video: TechSprint 2018 opening film

Developing solutions

Problem statements

Awards

Observing proofs of concepts

Find out more

Related content

Request an alternative format