“A regulator walks into a bank…”
While this might sound like the start of a bad joke among compliance officers sharing a drink after work, being “regulator ready” was a key message at the Accuity Client Summit held in New York in June.
In today’s climate, meeting regulatory obligations is more difficult than ever. This is due to a number of industry trends, including skyrocketing volumes of data that need to be screened, more regulations and regulatory scrutiny, the increasing complexity of sanctions, new markets (e.g., alternative investments, cryptocurrency exchanges) and smarter criminals.
Together, these trends form a potent cocktail of challenges that are driving the playbook of large and small organisations alike.
It’s no longer enough to have the right solutions, regulators want transparency, explainability, and proof that systems are doing what they are supposed to do. In fact, more than 60% of Accuity Client summit attendees indicated in a poll that “regulators asking for more proof” was one of the most challenging compliance pressures.
Whether screening for sanctions or to identify bad actors as part of an anti-money laundering (AML) program, it’s all about risk. Informative sessions explored the challenges from various viewpoints and discussed advancements in technology — and the Accuity solutions —that improve the detection, understanding, and management of risk.
Three key elements to mitigate risk in financial crime screening include:
- Detection – Reliably identifying risk is the first step for effective financial crime screening. Advanced tools such as artificial intelligence (AI) and machine learning – and their impact in redefining processes, workflow and customer experience – continue to be a hot topic. And with good reason, AI is transforming detection by dramatically improving accuracy and enhancing efficiency.Accuity recently released Firco Entity Resolution Filter, which uses AI techniques to deliver more accurate, relevant alerts with fewer false positives, improving the customer experience.
- Transparency – The days of a “black box” whose hidden algorithms magically spit out results is long past. Regulators want transparency to see how models work and how results were derived. But transparency alone is not enough – compliance teams must also understand and be prepared to explain the results as needed.Transparency and explainability are particularly critical with AI, where the decision flow may not be easily understood. However, the benefits of AI are so compelling that it is beginning to gain traction among regulators. The Financial Crimes Enforcement Network (FinCEN), the Office of the Comptroller of the Currency, and other agencies agreed “to encourage banks to consider, evaluate, and, where appropriate, responsibly implement innovative approaches to meet their Bank Secrecy Act/AML compliance obligations.”Technology solutions should be able to present scores or decisions in a human readable format. This might be as simple as highlighting key data elements contributing to a score, or as complex as describing interactive variables in an AI model.
- Provability – Once you understand how the model works, the final “frontier” is to demonstrate to regulators that the model is performing as expected. Testing and model validation reinforce the validity of the results. Tuning is important as well. Your software solution should provide flexibility to test and tune your system according to your organisation’s risk appetite.
Making AI Work
‘AI is not magic’, it requires good, clean data to perform properly and return meaningful results.
Gathering quality data is a well-documented challenge. Vendors like Accuity can enhance listings to improve the quality of data. However, organisations are advised to look at the lineage of data from the origin to the source to understand how it is truncated, enhanced and transmitted – and be prepared to explain it to regulators.
For successful implementation of AI solutions, organisations are best served by a measured approach:
Inventory – Do a data audit. Determine what data you have and how it is structured.
Define – Identify the problem you are trying to solve and what you wish to accomplish.
Identify – Decide which AI technique(s) can help and continue to experiment.
Bringing It All Together
In his keynote presentation on sanctions screening, guest speaker David Chenkin, managing partner at law firm Zeichner Ellman & Krause, discussed the importance of communication.
“Most regulators never worked in a bank and most technical folks never worked as a regulator, so a fundamental step to avoid misunderstanding is to get all stakeholders to the table,” said David.
There are opportunities to bridge the gap between the public and private sector, but the industry must come together and learn to speak the same language in order to collaborate. With everyone on the same page, it’s easier to find common ground, especially the next time a “regulator walks into a bank.”
Accuity Client Summits
The Accuity Client Summits, which are held each spring in New York, London and Singapore, are all about finding common ground. Bringing together client practitioners, consultants and subject matter experts in an educational forum enables us to continue to develop solutions that empower clients to meet the payments and regulatory compliance challenges they face every day.