The proliferation of new regulations in the financial sector after the Great Financial Crisis — in addition to firms’ own initiatives to improve compliance and reduce risk — have created a spike in both the demand for and cost of compliance and risk management talent. These fields are heavily reliant on manual processes that currently require armies of knowledge workers whose work, the safe navigation of complex and ever-developing requirements driven by an almost infinite matrix of factors, can never be as accurate as is ideal — indeed, Citi GPS estimates European and US banks have paid in excess of $150bn in litigation and conduct charges since 2011.
For financial institutions like HSBC and JPMorgan, where compliance and regulation staff account for 10% of operating costs ($270bn per year), AI provides a compelling alternative. This is despite the fact that AI itself comes with its share of regulatory burden. At the current state of art, artificial intelligence can build knowledge, understand natural language, interact naturally with human beings, overcome ambiguity and adapt to new types of problems and real-world workflows.
It is already being leveraged to disrupt multiple facets of financial enterprise: chatbots are interacting with customers, giving them better, quicker answers; in the back office, qualitative information, especially that mined from social media, is being folded in with quantitative information to achieve far superior risk analytics; and finance professionals are becoming disruptively productive by outsourcing their research to smart assistants.
In any of these fields and more, the premise is straightforward: By automating time-consuming low-to-mid value tasks, financial institutions can free employees to focus on more complex and value-added work. In terms of regulation compliance, there are three key ways AI can help: driving the identification and understanding of regulatory requirements, improving efficiency in addressing them, and dramatically reducing the risks of misconduct and error.
Take just one branch of regulations, Know Your Customer (KYC), and it’s easy to see AI’s transformative potential. For a financial institution looking to identify and verify the identity of its clients, an automated system could interact with and inspect the responses of clients, making its own judgments on whether those responses are satisfactory. It could also rapidly mine private and public repositories, assembling all information necessary for a knowledge worker wanting to better understand KYC, and how it relates to their organization. The overall result is a drastic reduction in the time taken to complete the KYC diligence process — from weeks to hours. This in turn means faster response, easier compliance and significantly better customer service.
KYC is the tip of the automating iceberg. From account openings and credit limit evaluations to the identification and explanation of changes in risk exposure, the use cases continue to expand.
- Know Your Customer
- Basel III
- SEC Filings
- Internal Compliance Reports
- Anti-Money Laundering
In the market of cognitive automation, two distinct models are emerging. An approach like one with IBM Watson takes an investment of time and money worthy of a supercomputer before combining text, speech, visual signals and others to solve foundational challenges. On the other end small companies take a more tactical approach, honing in on text-based workflows, going after specific problems like regulation compliance and achieving 98% accuracy within weeks. Many firms are building in-house teams to solve some of these problems in a bespoke way.
Irrespective of the approach, regulatory compliance and artificial intelligence are fast becoming wedded forever.