New Technology To Reshape Surveillance
New technologies such as machine learning and artificial intelligence will be actively utilised in surveillance in three years time according to the Association for Financial Markets in Europe’s second annual legal and compliance conference in London yesterday.
Stephen Strombelline, managing director at Capital Forensics, spoke on a panel about new technologies shaping surveillance and said it would take about three years for compliance functions to fully incorporate new techniques.
“Regulators have expectations on how firms use big data,” he said. “Random sampling will become antiquated and banks need to use natural language processing to identify key phrases in certain contexts.”
Paul Clulow-Phillips, global head of surveillance at Societe Generale, said on the panel that banks need to have a long-term plan on how to use big data and machine learning within surveillance.
“There is immense potential but banks are only at the start of the journey,” Clulow-Phillips added. “We also need to be able to explain to regulators how black boxes make decisions and to integrate new technology with monitoring of communication and trade data.”
Clulow-Phillips continued that Societe Generale is in the first year of a five-year plan to incorporate new technologies in compliance.
AFME and EY also published a report today considering compliance within wholesale investment banks and the potential challenges it will face.
— AFME (@AFME_EU) October 2, 2018
Stuart Crotaz, financial services partner at EY, said in a statement: “The speed of change and the increasing importance of data and analytics means that compliance needs to keep evolving.”
The report said compliance reporting is still largely manual and so takes resources and time. However, compliance needs to keep up with the increasing use of automated trading platforms, algorithmic trading, social media, and the more complex transactions and structures being developed by the business.
The report added that in order to move to a different operating model there should be a shift towards compliance being a data user, not a data generator.
“Fundamentally, compliance should take a step back and look to use and leverage all streams of data (transactional, behavioural and social) to identify risks and act as an independent overseer and advisor,” said the study.
James Kemp, managing Director at AFME, said in a statement that expectations for compliance functions have never been higher, against a backdrop of significant regulatory change and conduct issues.
Kemp said: “Where previously advising on regulatory requirements and monitoring adherence to company policies formed the core of a compliance officer’s role, teams are also increasingly expected to take a more strategic and proactive role in anticipating and managing risk.”
Structured data will push the platforms beyond their pre-investment tasks.
Legislators need to pay attention to results instead of the methods.
The new offering supports reconciliation, matching and exception management applications.
Method reuses a model developed for a task as the starting point for the next task.
A strong data foundation, dedicated AI teams, and C-suite buy-in are needed for success.