How reducing risk is a delicate dance on a regulatory tightrope
Regulations are the music to which businesses perform a delicate dance. If you took to the stage, a move or two out of sync would not cause the floor to drop out from underneath your steps. But for businesses moving to the rhythm of regulations, most particularly for banks and other financial institutions, the consequences of breaking tempo with compliance can be catastrophic. Who can stand the pressure?
In this moment of global upheaval, the financial services sector has never been under more regulatory pressure. COVID-19 rages. An increasingly volatile world economy hangs overhead. Regulatory bodies have only constricted tighter, requiring financial organizations to perform more complicated choreography to keep up with changing laws and rules governing their industry. All the while organizations are still finding their feet with emerging RegTech and bringing their swelling information estates to bear as required under regulations, but also as a fuel for innovative AI solutions. With pressure mounting, how can financial institutions navigate the regulatory tightrope?
In times of crisis, the best people to turn to for guidance are the veterans of the industry. Sébastien Piednoir is the Chief Compliance of CACEIS. He began his career with Credit Agricole Group in 1994 as a research and development engineer. Rising quickly through management, Piednoir became Global Head of Internal Control of Capital Markets in 2004, Chief Operating Officer of Capital Markets in New York four years later, and by 2013, he created the position of owner of the Clients Management processing. With more than a quarter-century of financial services experience, Piednoir provides valuable counsel on how financial institutions can enhance knowledge workers with trusted AI to relieve the regulatory pressure.
Watch a 0:27 summary above of a leadership conversation on compliance between Sébastien Piednoir from CACEIS and Jean-Philippe Desbiolles from IBM, and watch a video of the full conversation (19:34) here.
In an era of endless data, compliance decisions get harder
If only the simple compliance advice “follow the law” was enough to resolve a financial institution’s regulatory requirements. But as S. Piednoir confesses compliance decisions are seldom this straightforward.
“Sometimes it’s quite easy; you don’t need a compliance officer. You just need to follow the rule and the law,” he said. However, according to S. Piednoir, most of the time compliance is a question of interpretation: for example, do we work with one group or another based on a perception of non-compliance risk.
Extracting insights from unstructured data is the basis to determine these compliance interpretations and make a good decision, but it comes with a major caveat. “You need to make a decision that makes sense and is in the interest of the client. It needs to comply with the current law, comply with future law. But at the same time the data that you have access to is going to get bigger and bigger every year,” said Piednoir.
From his perspective, a major problem happening today is that people are spending more and more time accessing large volumes of data and transforming it, while not spending enough time to make complex decisions.
“This is generating a stressful situation,” said Piednoir. “Because the decision is stressful, it leads to even more time formularizing the data—rather than focusing on what is a good decision. We need to help humans spend less time in accessing and transforming unstructured data to structured data.”
“This is the domain of AI,” said Piednoir.
Trusted AI empowers the compliance professional
People think that if they scale artificial intelligence they will lose the “know-how” of what is key in their business. According to Piednoir, it is in fact exactly the opposite.
Take the example of company financial reports. These 100-300 page documents are a key way for banks and financial institutions to evaluate the risk of doing business with a particular organization. But professionals would spend too much time combing through these 100-300-page financial report documents. Now these manual tasks can be automated. Built on IBM Watson, Crédit Agricole employs a tool that reads 300-page documents in 5 minutes.
This is the “management of the know-how,” said Piednoir. Knowledge workers bring their expertise to train AI. They set the controls. They become masters of what is produced by the AI model.
“In a sense, you train the computer just as you train another human,” he said. “I cannot buy a blackbox. But I am happy to participate in an added value product.”
“This is a question of partnership,” said Piednoir.