AI Dives into Compliance
For financial firms, the back office faces a similar challenge digesting, analyzing, and discovering the answers that it wants from the ever-growing amount of data as the front office.
Artificial intelligence-based platforms that use data analytics give compliance departments an alternative to throwing a conference room full of lawyers to review data manually.
It is not a panacea, according to Wayne Matus, managing director, group investigations at UBS and who spoke on an AI panel hosted by Exiger.
“There is a surge of the quantity and quality of data, which let us get to the truth much more accurately for the firm time,” he said. “There is no substitute for the person who can see the pattern that the machine cannot.”
Using data analytics powered by AI, however, can provide compliance departments with views that they never had before.
Firms can triangulate the truth more exactly by knitting together data sources that firm would never have in the past, such as data from trading systems, customer relationship management platforms, human resources platforms, and outside social media, said Dan Adamson, President, DDIQ and Global Head of Cognitive Computing, Exiger Analytics.
“It can make a much stronger case than before,” he said. “The amount of behavioral science done around people is greater than any time before.”
One investigation by an unnamed firm into a suspected trader’s behavior was found not in the trader’s email or trading blotter but a call to a third-party help desk, according to Brandon Daniels, managing director & president at Exiger. “Why would I ever think a smoking gun would have been a help desk conversation?” he asked.
“Analytics help us see the whole when we only saw the pieces previously,” added Adamson,
The only major limitation to the scope and quality of data sets are the legacy systems that each firm has implemented, noted fellow panelist Alexander Southwell, a partner at Gibson Dunn & Crutcher.
UBS’ Matus again voiced his concerned that the pendulum might be moving in the wrong direction as firms try to replicate only portions of what seasoned compliance staff brings to the table. Instead, Matus suggested that individuals take a more personal approach to learning how to apply AI-based tools within the compliance field.
“London taxi drivers have ‘the knowledge’, but how do I get it?” he asked. “I could hire someone with it to teach me, but I say get involved. Find the places where it is being used in the trenches and then start getting the knowledge.”
Deep learning may do to HFT what HFT did to statistical arbitrage.
AIs may flag "unambiguous" rule violations in the future.
A difficult, but not impossible, task
The firm completes natural language processing PoC.
Next-gen advisors broaden product support and depth.